Itujfa 2018 1 by LookUs Scientific

ITU A|Z • Vol 15 No 1 • March 2018

Contents Y. Çağatay Seçkin • Editor Editorial

I-II

Dossier: Future trajectories of computation in design Ethem Gürer Dossier Editorial

1-3

Elif Ensari, Bilge Kobaş Web scraping and mapping urban data to support urban design decisions

5-21

Pınar Çalışır Adem, Gülen Çağdaş Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

23-38

Miray Baş Yıldırım A model for parameterization of urban regulations

39-51

Ömer Halil Çavuşoğlu, Gülen Çağdaş Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

53-64

Şeymanur Yıldırım An investigation on algorithm aided BIM approaches to increase collaboration and optimisation in project phase: A case study 65-77 Dagmar Reinhardt The sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance Güzden Varinlioğlu, Burkay Pasin, Hugh David Clarke Unconventional formulations in architectural curricula: An atelier on design for outer space architecture Orkan Zeynel Güzelci, Sinan Mert Şener A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

79-92

93-105

107-122

ITU A|Z • Vol 15 No 1 • March 2018

Theory Ceyda Sarıca, Ebru Çubukçu Evaluating color combinations using abstract graphics versus pictures of simulated urban settings Meriç Demir Kahraman, Burak Pak, Kris Scheerlinck Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

123-134

135-145

Berfu Güley Gören, Lale Berköz Accessibility of transfer centers with different transportation modes for disabled individuals

147-161

Ümit Arpacıoğlu, Mustafa Özgünler An evaluation on immaterialisation phenomenon in religious spaces of architecture

163-175

Editorial Y. Çağatay SEÇKİN • Editor Another year has passed, 2017 has brought many design innovations and we are full speed ahead in 2018. The year 2018 in architecture and design, as you might expect, will involve significant events, new designs and technologies, for sure. Especially in digital practice and technology, virtual reality (VR), augmented reality (AR), and real-time rendering will continue to transform the way we design. Last year, we have seen much excitement about blockchain, particularly around cryptocurrencies, or digital currencies, like Bitcoin. The newly emerging decentralized technology, blockchain, has the undeniable potential of being the most disruptive technology in the 21st century, after the internet. A blockchain is a non-corruptible account book that keeps a record of all transactions that take place across a peerto-peer network. The incorporation of blockchain networks into the design and construction industry will have a profound impact, facilitating quicker and more secure transactions. No wonder, developing technologies will continue to give us the opportunity to gain new perspectives that allow us to design in a more dynamic manner than ever before. Depending on the case and expectations for 2018, the dossier of this issue has been titled as “Future trajectories of computation in design” which is edited by Ethem Gürer, PhD. The articles of this dossier were selected from CAADFutures 2017 held on July 10-14, 2017. The dossier has the articles as follows: “Web scraping and mapping urban data to support urban design decisions” by Ensari and Kobaş, “Interpretation of urban data in the complex pattern of traditional city: The case of Amasya” by Çalışır Adem and Çağdaş, “A model for parameterization of urban regulations” by Baş Yıldırım, “Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions” by Çavuşoğlu and Çağdaş, “Investigation on algorithm aided BIM

standards to increase collaboration and optimisation in project phase: A case study” by Yıldırım, “Sound of space in 3 robotic prototypes: Introducing 6axis robotic fabrication to shape macro-and micro-geometries for acoustic performance” by Reinhardt, “Unconventional formulations in architectural curricula: An atelier on design for outer space architecture” by Varinlioğlu, Pasin and Clarke, “A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings” by Güzelci and Şener. In addition, this issue has five articles in the theory section. Ceyda Sarıca and Ebru Çubukçu wrote the first article. Their article entitled “Evaluating color combinations using abstract graphics versus pictures of simulated urban settings”. According to the paper, colors in urban environments influence people’s judgments of environmental quality; and color combinations have received remarkably little empirical attention and no study compared people’s responses to ‘abstract color combinations’ and ‘color combinations in urban settings’. This study aims to fill this gap by both investigating people’s preference for various color combinations and comparing people’s evaluations of “abstract color compositions” and “contextualized color compositions pictures of simulated urban settings”. The second article in the theory section, “Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach”, written by Meriç Demir Kahraman, Burak Pak and Kris Scheerlinck, aims to anatomize the paradox of public space from also the insights of social sciences in the conditions of representative democracy. As the main contribution of this study, authors introduce a re-interpretation of Lefebvre’s multi-triads and operationalize his concept of heterotopia to offer a deeper understanding in revealing the paradoxical production of public spaces. They conclude that the social production of a heterotopia is the manifestational realization of an ideal public space and the dissolution its paradox for only a temporary period of time.

Berfu Güley Gören and Lale Berköz try to create a new concept on the integration of universal design principles and the criteria of accessibility for disabled people in urban areas, in their paper entitled “Accessibility of transfer centers with different transportation modes for disabled individuals”. Therewith this new composition would be included to the disability research literature. Furthermore, in this study a new matrix to calculate the accessibility score of transfer centers has been developed. The last paper of this issue is “An evaluation on immaterialization phenomenon in religious spaces of architecture”. In this study, Ümit Arpacıoğlu and Mustafa Özgünler develops a different approach for the evaluation of religious space in terms of use of material and its religious expression. The paper emphasizes the two contemporary concepts that are quoted from art through the examples of two monotheistic religions: Islam and Christianity. Lastly, before I end this editorial, let me tell you a few words about Aydın Boysan who died on January 7, at the

age of 96. Boysan studied architecture in Academy of Fine Arts - later renamed as Mimar Sinan Fine Arts University. He served as an architect more than 70 year. During this period, he won many national and international architectural design competitions. The total area of his building designs was about 1,500,000 square meters. In 1954, he became the first secretary general of the Chamber of Architects, and he taught in İstanbul Technical University, between the years of 1957 and 1972. As his students and colleagues, we acknowledge his impact, and we remember him with respect and gratitude for all he gave to us as our master and teacher. As it always has been, I would like to thank all our readers for the support they provide to the Journal. We really look forward your comments, contributions, suggestions and criticisms. Please do not hesitate to share with us your feelings and especially, let us know if you have ideas or topics that we could be focusing on. Enjoy your reading and meet with us again in next issue on July 2018.

Dossier Editorial: Future trajectories of computation in design Ethem GÜRER Computation has become a widely embraced notion in architectural design. While designers traditionally rely on intuition and experience to deal with design problems, computational approaches aim to enhance that process by supporting abstraction, algorithmic and parametric thinking, collaborative design, form studies, complex modeling, automation, simulation and fabrication. In this regard, computational design not only alters the very nature of design activity (thinking and doing), from intuition to logic, as well as from product to process, but also offer powerful media for the communication of different disciplines, to the designer in the multidisciplinary world of the 21st century. The dossier subject for A|Z ITU Journal of the Faculty of Architecture 15 (1) special issue is entitled “Future Trajectories of Computation in Design”, as parallel to the conference CAADFutures 2017 organized in Istanbul Technical University, Taşkışla. The main purpose of this special issue is to question and to discuss how the practice of design is reforming (and will reform) itself in response to these crucial changes. The dossier comprises eight articles, which introduce not only the state-of-the-art in the computational design field, but also recently developed computational tools, technologies, methods and implementations over specific case studies of different design context and scale. The three papers that were originally presented at the CAADFutures 2017 Conference in Istanbul went through a rigorous double-blind- refereeing process, and benefited from discussions during the conference. The remainder of the papers arrived through the conventional journal process upon invitation and underwent the same blind peer review by experts. Departing from the fact that modern cities generate data in increasing speed, volume and variety which is

more easily accessed and processed by the advance of technology, Elif Ensari and Bilge Kobaş present methods in order to collect and visually represent these data with the aim of supporting urban design decisions. The authors propose to use web scraping methods to gather a variety of publicly available data within the Kadıköy municipal boundaries of Istanbul, as well as a visual programming software to map and visualize this information. Urban conditions such as demographic and economic trends based on online real estate listings, spatial distribution and accessibility of public and commercial resources are discussed through the presented method and superposition of the resulting maps. To comprehend the inner nature of complex traditional cities, Pınar Çalışır Adem and Gülen Çağdaş introduce Data Mining algorithms of Geographic Information System (GIS) technologies as a holistic data gathering and processing approach. The paper firstly discusses the risks of uncontrolled urbanization activities like top-down planning decisions and designer’s personal motives over the local character of traditional cities. Through the case of Hatuniye Neighborhood in Amasya, the authors then present the logic of their interpretative method for dealing with traditional urban data, which is mainly based on topographical and morphological attributes, in three consecutive phases: formation, analyses and interpretation of data. Referring to the three phases of the established methodology, Adem and Çağdaş emphasize the importance of the use of computational models and approaches in collecting and interpreting local urban data to sustain the special character of traditional environments. The article which is titled “A Model for Parameterization of Urban Regulations” by Miray Baş Yıldırım broadly focuses on the significance of multidisciplinary approaches and participatory processes in the urban regeneration processes. While the first section presents a brief summary of the parametric regulation modeling and design guides within the context of urban design problems, the following parts describe the generation of the model as

a parametric regulation modeling tool for design decision support. By using real-time regulations and standards as fundamental design data, the proposed model allows designers to generate and test various design scenarios. Accompanied with some technical limitations, the model differs from others by customization of local regulations for evaluating design alternatives. The transformation of textual regulations and standards into computational design rules is the main challenge of the study. Ömer Halil Çavuşoğlu and Gülen Çağdaş discuss opportunities of Building Information Modeling (BIM) from an unfamiliar perspective that aims to integrate BIM tools into early design stages. With the use of different data gathering methods such as quantitative ones (questionnaire), qualitative ones (pure observation, participant observation, in-depth interviews and focus groups) and protocol analysis (retrospective analysis) through five case studies that were implemented with 25 unique participants, the article elaborates five crucial impacts for the BIM environment that contribute to the early design stages: design exploration, 3D modeling, parametric modeling, conceptual energy modeling and decision support system. Similar to the previous article, Şeymanur Yıldırım draws attention to the increasing role of BIM in design, but this time from an algorithmic perspective and in an architectural project development context. The lack of healthy communication between different disciplines and actors, and the difficulty of managing big data in a design process are introduced as the major problems of BIM for architecture offices. Through a specific case study, the paper not only challenges these problems but also offers highlights for the companies that intend to implement algorithm aided approaches in BIM during project design, development, management and construction processes. In order to shape both macro and micro geometries for acoustic performance, Dagmar Reinhardt introduces 6-axis robotic fabrication which is based on a range of mathematical equations that regulate physical properties of spatial surfaces and pattern

details through the implementation of three robotic prototypes. By interfacing computational scripts, mathematical source codes, and structural analysis, this exhaustive research discusses the opportunity of producing a compelling and distinctive topography of multiple scales and dimensional geometries for shaping sound in the context of performance spaces and theatre acoustics. From a broader point of view, Güzden Varinlioğlu, Burkay Pasin and Hugh David Clarke shed lights on the problem of integrating digital methods and tools into the architectural curriculum through a case study of a workshop entitled “Mission Mars 2024: A Biomimetic Structural Organism”, as part of the studio course ARCH 202 in the spring semester of 2017 at Izmir University of Economics Department of Architecture. As demystified by the authors, the paper presents a critical approach to understanding the impact of digital tools and methods on the learning outcomes of the students, which are, discussed and demonstrated based on four studio outcomes. Orkan Zeynel Güzelci and Sinan Mert Şener introduce the concept of entropy as an objective, precise and quantitative methodology for measuring various types of information embedded in the built environment or buildings. For this purpose, the authors investigate 24 municipality building projects which were designed for architectural project competition between 2015 and 2016 in Turkey, and which have received awards. In order to reveal objective similarities, 24 projects were re-evaluated by calculating five different entropy values. The study differs from the precedents by using multiple interrelated entropy values for the evaluation of a specific architectural typology. The presented algorithm is also claimed to be a model for making some predictions about the potential of project winning a prize in a competition. As the guest editor of this special issue on Future Trajectories of Computation in Design, I would like to express my sincere gratitude to Elif Ensari, Bilge Kobaş, Pınar Çalışır Adem, Gülen Çağdaş, Miray Baş Yıldırım, Ömer Halil Çavuşoğlu, Şeymanur Yıldırım, Dagmar Reinhardt, Güzden

Varinlioğlu, Burkay Pasin, Hugh David Clarke, Orkan Zeynel Güzelci and Sinan Mert Şener for their valuable contributions to A|Z ITU Journal of the Faculty of Architecture 15 (1) special issue. I am also grateful to each re-

viewer for their guidance accompanied with assiduous reviews. Moreover, I would like to thank the editors of the journal for such an opportunity of discussing the horizons of computation in design through relevant studies.

ITU A|Z • Vol 15 No 1 • March 2018 • 5-21

Web scraping and mapping urban data to support urban design decisions

Elif ENSARİ1, Bilge KOBAŞ2 1 elif.ensari@gmail.com • Architectural Design Computing Program, Istanbul Technical University, Istanbul, Turkey 2 bilge@super-eight.com • Bits’n Bricks Analytical Research Center, Istanbul, Turkey

doi: 10.5505/itujfa.2018.40360

Received: October 2017 • Final Acceptance: November 2017

Abstract Cities generate data in increasing speed, volume and variety which is more easily accessed and processed by the advance of technology every day. Consequently, the potential for this data to feedback into the city to improve living conditions and efficiency of utilizing resources grows. Departing from this potential, this paper presents a study that proposes methods to collect and visualize urban data with the aim of supporting urban design decisions. We employed web scraping techniques to collect a variety of publicly available data within the Kadıköy municipal boundaries of Istanbul and utilized a visual programming software to map and visualize this information. Through this method and superposition of our resulting maps, we visually communicate urban conditions including demographic and economic trends based on online real estate listings as well as spatial distribution and accessibility of public and commercial resources. We propose that this method and resulting visualizations present valuable potential in supporting urban design decision-making processes. Keywords Geospatial data, Urban data, Urban design decision support, Web scraping.

1. Introduction Data turns out to be more relevant for urban design as it becomes more ubiquitously available and accessible through the advance of technology and as it is more widely incorporated in the decision-making processes regarding the urban environment. Recent years have seen the proliferation of big data as well as research concerned with its integration into the analysis, management and design of cities. Numerous academic, civic and commercial research centers have been founded dedicated to studying the urban environment through information generated by the cities (Batty, 2013). While the scale, variety and production speed of data (3Vs) increase every day, its availability, accessibility and communication to designers, decision makers and to general public remain a critical issue. This is due to the range of skills required to access and process the data in order to make it relevant and meaningful. The aim of our study is to present an urban data collection and visualization methodology which can help urban designers produce maps to communicate various urban patterns to stake holders and decision makers in order to support urban design decisions, as well as to communities to better inform participatory processes. We also present a set of maps produced following this workflow, along with some basic geospatial analyses we have performed to create them. The questions we take on to answer through the study presented in this paper are as follows. • How can sources of publicly available data with geolocation information be relevant and useful to urban designers? • How can this data be gathered and visualized? • What kind of knowledge can be generated through the visualization of this data that may be presentable to decision makers even before any geostatistical analysis is carried out? The study is unique in that we present a methodology to compile a rich dataset which we propose to be relevant for urban design decision making processes, consisting of information drawn from multiple sources rather than studying a single source of data

and its relevance for a specific urban issue. We also present visualization of each data set with some basic spatial analysis. Through this, we aim to aid quick and easy visual analysis and better communication of this data to designers and decision makers. Furthermore, Kadıköy region of Istanbul has never been subject to such a comprehensive study in terms of the variety of urban data and data sources utilized as well as the techniques adopted for mapping and visualization. The data collection was done utilizing Python scripting and all mapping, visualization and visual analysis processes were done using the visual programing environment Grasshopper for Rhino3d. Our research has revealed various urban patterns that became apparent through the maps we have created utilizing the data collected from multiple geolocated urban data sources. Two most significant cases we have observed to have potential in contributing in the urban design decision dialogues are the exposition of public and commercial resource distribution, and demographic and economic trends read through the various maps we present in the following sections. We aim to develop this study by integrating more comprehensive analyses to our workflow in future research. 2. Literature review While cities around the world are growing in size and population every day, digital devices and systems are becoming more and more integral to their functioning in parallel with technological advancements. This results in the generation of massive amounts of data in the urban environment which have already been subject to a large body of research motivated with the belief that cities can learn from this data and become ‘smarter’ in order to provide better living conditions for their residents while utilizing resources more efficiently (Batty et al., 2012; Townsend, 2013). Mobility patterns of urban dwellers drawn from digital public transportation ticketing systems, cellular phones or cameras detecting traffic patterns, GIS data compiled and made available to public by governments, various use and activity data shared through

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

personal mobile devices and publicly available social media content are some of many examples of such data. A part of these studies in urban data focuses on gathering, analyzing and visualizing data which is inherently linked to geolocation information, namely geospatial big data (Lee & Kang, 2015), making it possible to map and observe geospatial patterns of urban phenomenon. While such studies often bring together geographers and computer scientists based on interest and technical skills required; urban planners, designers and architects also see value in this data which can provide precious insight on the functioning of cities and support urban decisions in order to improve the quality of the urban environment (Glaeser, Duke-Kominers, Luca, & Nalk, 2015). One of the most popular subjects of study of dynamic urban data is real estate rental and sales prices, mapping and geographical analysis of which allow for understanding location driven contributors to property values (Cohen & Coughlin, 2008; Geoghegan, Wainger, & Bockstael, 1997; Waddell, Berry, & Hoch, 1993). More recently, studies utilize housing price information from public real-estate websites (Bency, Rallapalli, Ganti, & Srivatsa, 2017; Boeing & Waddell, 2016), facilitating the use of very large and up to date datasets, which allows for the development of more accurate housing price prediction models. One such study finds correlations of housing prices with geo-tagged social media content (Li, Ye, Lee, Gong, & Qin, 2017), which is another type of data that has been subject to urban analysis research as an indicator of urban activity (Jenkins, Croitoru, Crooks, & Stefanidis, 2016; Schreck & Keim, 2013). Walking, running and other forms of exercise data tracked and shared by urban residents have also been studied to better understand qualities of preferred locations and routes for recreational activity in the urban environment (Balaban & Tuncer, 2016; Clarke & Steele, 2011). There are various methods and techniques which allow for management of big data, yet means of its collection and visualization is the main subject of our interest in this research. As a method

to draw data from public web sites, web scraping which we utilize in our study involves automated means to query and download large sets of geospatial data. Boeing and Waddell (2016) utilize this technique to analyze real-estate listings in the US and are able to compare eleven million listings all over the country with government declared fair market rents in 58 metropolitan areas. APIs (application programming interfaces) provided by many online platforms such as Google, Instagram, Facebook, Flickr, Foursquare or Twitter make it possible to query and download user generated content, which becomes interesting for urban researchers when the data is accessed together with geolocation information. Among several recent studies that utilize social network data to better understand urban patterns, Cranshaw and colleagues (2012) detect areas of concentrated urban activity (“Livehoods”) using Foursquare check-ins and Jiang, Alves, Rodrigues, Ferreira, & Pereira (2015) estimate land use based on Yahoo’s point of interest (POI) data with a higher accuracy than with traditional methods. Chen (2014) provides a good overview of data collecting methods from social media platforms including web scraping and the use of APIs that were also utilized in our study. All geospatial data is already big data (Lee & Kang, 2015), and velocity, variety and volume are the 3Vs considered to define big data. While collecting, managing and analyzing such rapidly generated large scales of data already pose a significant challenge, visualizing and communicating it is utmost critical in deriving meaning and value from geospatial data in order for cities to learn from it. There are several tools from geographical information systems (GIS), building information models to virtual simulation environments that make available various visualization techniques (Pettit et al., 2012) for geospatial data. A project that demonstrates the power of visualization of geospatial data, “Million Dollar Blocks” (Kurgan, 2013) maps the home addresses of criminals as they are admitted into prison as well as the length of their stay in five of US’s largest cities and reveals

Web scraping and mapping urban data to support urban design decisions

Figure 1. The workflow.

a remarkable concentration of these addresses around a few city blocks. One of the project’s most striking visuals maps the money spent by the state on each criminal to their home address, displaying 4.4 million dollars spent on criminals coming from within just four blocks in Brownsville, Brooklyn. A study conducted in a newly activating area of Amsterdam, The Knowledge Mile (Niederer, Colombo, Mauri, & Azzi, 2015), presents geographical maps of companies and their online hyperlinks, most shared images and most shared locations using data gathered from online platforms Google Search, Panoramio, Instagram and Foursquare. Making apparent the online presence and character of the urban axis in the focus of the research, these maps were later shared with local stakeholders in participatory design sessions and reportedly initiated conversation and further field work. Having various similarities with the research presented above, our study presents multiple layers of geographically linked urban data drawn from dynamic sources of public web content, mapped to reveal patterns of urban activity and distribution of commercial and public facilities. We propose that

these visualizations created through basic spatial analysis and mapping techniques allow for human visual assessment of a rich set of urban data, which is deemed as valuable as automated analytical processes (Schreck & Keim, 2013). 3. Methodology Our process involves the following steps: drawing of data from online sources, organizing the data, performing mathematical operations to calculate some additional values on the database, mapping the data based on geo location information, performing basic geometrical analysis and visualization. Our workflow is presented in Figure 1. We utilized multiple sources of publicly available data from within the Kadıköy municipal boundaries of Istanbul, to draw already geo-tagged data or data with address information that we converted to latitude and longitude values through Google’s Geocode API. While mapping and visualizing the raw data we collected, we applied different levels of intervention; from simply positioning the circles at geographical locations and scaling them based on their values within the domain of the data set or connecting the geolocation

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

Figure 2. Unit rental prices (rent price per sqm).

points to create routes and superposing them, to performing surface to surface intersections to calculate areas of influence or accessibility of amenities automated through custom codes. Our data comprises real estate rental and sales prices, property square meters, property bedroom numbers, ages of buildings, uses of buildings, social media activity, public park boundaries, commercial facilities, public transportation and other public facilities. For the data collection phase, we use Python programming language to send requests composed by the Postman to the source web siteâ&#x20AC;&#x2122;s servers and generally download the data in Json format. We then convert the data to csv (comma separated value) format and import it to Grasshopper visual programming environment, an add-on for Rhino3d, using Meerkat plugin. Here we process the values within the database or run simple spatial analysis after converting them to geometrical entities. All the steps carried out in the Grasshopper environment besides the final graphic touch-ups are automated so that they are easily repeatable for different locations in future studies. Below we present the details of our methods grouped based on the sources

of data and the themes of the maps that were created. 3.1. Housing and real estate We utilized one of the most widely used real estate listings websites in Turkey with 42.6 million active members visiting the portal 263 million times per month that allows for posting and managing of apartment rental and sales listings by both the owner of the property as well as real estate agents. The website has a map interface through which we were able to draw geolocation, size (in sqms) number of rooms and comments of the listing owner of each apartment in the Json format. Then we sent a request to the page of each listing to draw further text based information including the type of heating, additional amenities and whether the listing was posted by the owner or the real estate agent. The text data was parsed using the Beautiful Soup Library for Python and converted into csv format with Csv Library for Python. The csv files were then imported to Grasshopper and processed to calculate the rental and sales prices per sqm and create maps (Figures 2 & 3). In these maps, the radius of each circle positioned at the geolocation of the

Web scraping and mapping urban data to support urban design decisions

Figure 3. Sales prices per sqm.

Figure 4. Number of bedrooms for each house listing (rental and for sale).

listing is scaled proportionately with its price divided by sqms, after the outliers were excluded from the dataset. The number of rooms per apartment and building age data were combined

from both rental and sales listings and mapped with colors (Figures 4 & 5). Payback values were calculated per 100m by 100m area based on the average apartment rental and sales prices

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

Figure 5. Ages of the buildings of the listings (rental and for sale).

Figure 6. Payback time.

per sqm available in each grid square (Figure 6). A GIS file created by Istanbul Metropolitan Municipality was used to map the residential/non-residential build-

ing use information, with the building footprints colored using a gradient based on the proportion of the number of non-residential units to the residential ones for each building (Figure 7).

Web scraping and mapping urban data to support urban design decisions

Figure 7. Residential and non-residential uses in buildings.

Figure 8. Circle grid system for Instagram queries.

3.2. Social media Instagram was the source of social media data we utilized for this study as it is known to be favored over Twitter and other forms of social media in Tur-

key, with 16.34 million users by 2016 (Statista.com, 2016). Python code and Instagram’s own API was used to query 100 posts at a time, from within 150 m radii of 159 points arranged to cover all

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

Figure 9. Cafes, Instagram posts and their like numbers.

the area within the municipal boundaries of Kadıköy (Figure 8). Besides the geolocation data, number of likes and the text accompanying each image was drawn. The duplicate posts downloaded due to inevitable overlaps of circles were cleaned up. Posts that were liked over 1200 times were counted as 1200 due to their low frequency and posts within a distance of 5m’s to each other were combined by adding up their number of likes for the sake of a more legible representation. A map was created with the scale of each circle representing the like numbers per post point (Figure 9). The text data drawn from the Instagram posts were analyzed using a custom definition separating each word using Python split function defining space sign “ “ as a delimiter and hashtags were appended to the data of each post in the csv file. Then the frequency of each hashtag was calculated per 100m by 100m grid and mapped (Figure 10). 3.3. Public and commercial amenities Using the Google Places API, we drew the geolocations and ratings of Cafés and Restaurants which were marked on Google Maps. Similarly

with the method we drew Instagram posts, we queried 159 points with 150m radii and recorded 60 places at a time. Then we cleaned up the duplicates and created a simple point-location map. One of our resources was a widely used food delivery portal, Yemeksepeti.com, with over 5.2 million users by 2016 (2017), that directs orders from users to member restaurants in Turkey. Each item on a restaurant’s menu is displayed with its price, however, the restaurant’s address is not provided on the website. First, we queried the geolocation information of each restaurant using Google’s Geocode API. Then the prices of the most commonly sold items which were a 330ml can of Coke and 250ml cup of Ayran (a yogurt based Turkish beverage) were queried. We also drew the hours of operation for each restaurant. The website does not provide the menu items in a database format therefore a custom Python web scraper script was developed which searches for all restaurants delivering to the area and then follows each of their links to access their menus. Beautiful Soup Library was used to convert the HTML formatted site into a formatted spreadsheet, the prices were converted to the scales of circle radii distributed

Web scraping and mapping urban data to support urban design decisions

Figure 10. Hashtag map.

on the map based on the geolocation of each restaurant and the number of operating hours were converted to color codes to be visualized on the map (Figures 11&12). 3.4. Transportation and access We drew the geolocation data of stops for buses and metrobuses (a rapid bus transit line using dedicated bus lanes for much of the routes) from IETT’s (the general transportation department of the City of Istanbul) website as well as the number of times a bus or metrobus passes through each stop to create routes on a map (Figure 13). We used their frequency data to define the thickness of the route lines on our map. Then for each building centroid, we computed the number of public transport seats accessible within 400m, 800m and 1200m radii. For this, we

calculated the number of stops within these circles first, and then multiplied the seated passenger capacity of each vehicle with its frequency throughout the day. Without drawing the route lines, we did a similar calculation for ferries and subway lines, and added up all the seated passenger capacities within the mentioned radii distance from each building centroid. We illustrated these numbers through a gradient of colors for each building, representing a scale of accessibility to public transportation. (Figure 14) Another map we created aims to visualize the accessibility of recreational green spaces from buildings. For this, we utilized the geographically located outlines of public parks drawn from the GIS files and made surface intersections of them with the circles of 400m, 800m and 1200m radii from

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

Figure 11. Soft drink prices.

Figure 12. Operating hours of the restaurants.

each building centroid. The areas of intersecting regions gave us a domain of accessible green space sqms that we mapped as a gradient of colors for each building polygon (Figure 15).

We drew the exercise routes of KadÄąkĂśy residents from an activity tracking social network that allows for sharing of walking, running, biking and swimming activity data. Superposing these routes revealed

Web scraping and mapping urban data to support urban design decisions

Figure 13. Bus and metrobus lines.

Figure 14. Accessibility to public transportation seats.

the most popular streets and in one case a swimming route, as well as the approximately estimated geographic locations of the buildings, residents of which use these routes most frequently (Figure 16).

4. Observations and urban design decision support We propose that the set of maps produced by our automated workflow are powerful in demonstrating how urban

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

Figure 15. Accessibility to green spaces.

Figure 16. Exercise routes.

data can be basically organized and visualized to communicate meaningful observations and aid urban design decisions. Although the quantitative results we aim to generate in future studies

through geospatial analysis of this data will advance this research significantly, we believe that the current stage of the study we present here is as useful in revealing the relevance of geospatial

Web scraping and mapping urban data to support urban design decisions

information for urban design. Some of our observations at this stage of our research are presented below. We also present ways in which the visual information revealed in these maps can support urban planning and design decisions following each of our observations. There is a concentration of apartments for sale and higher sales prices along the coast and a concentration of rentals of 1+0 and 1+1 units on the west of Söğütlü Çeşme (Figures 2, 3 & 4). This is informative about the demographics of residents around these neighborhoods and can guide the municipality in deciding on the kind of investments they should make, may they be introducing co-working spaces, libraries, professional education facilities, kindergartens or retirement homes. Our Instagram hashtags and soft drink prices maps (Figures 10 & 11) similarly contribute in revealing the demographics of the local residents and visitors, as well as providing a glimpse of their consumer trends. Urban transformation which is encouraged by the policy to renew non-earthquake resistant housing stock of Istanbul is made apparent in our maps displaying the ages of buildings (Figure 5), showing the geographic distribution of recent construction. The rapid transformation within the Caddebostan area in the last couple of years is revealed which may be obvious to the neighborhood residents but no known maps exist with updated information available to public. The local authorities can utilize this information not only for understanding areas where transformation is slow, but also where it is rapid, and they can work with urban designers to take precautions against disadvantageous consequences of such a high density of construction sites and the rapid change of the built environment for the local residents. Together with the cafes & Instagram likes map (Figure 9), our residential-nonresidential buildings map reveal (Figure 7) the main commercial axes and the mixed-use pattern in Kadıköy which is known to be an indicator influencing the walkability in a neighborhood. The operating hours of restaurants (Figure 12) influence

the liveliness and feeling of safety on a street, contributing to the “Eyes on the street,” a concept first proposed by Jane Jacobs (1961). The geographical distribution of commercial amenities and social media activity can inform urban decisions of zoning and investment in new facilities or services by revealing if the existing facilities are sufficient or attractive to urban dwellers and how commercially active a street is throughout the day. Areas that are lacking in such amenities can be prioritized for making changes in planning to encourage a more vibrant street life and commercial activity on their streets. Our public transportation and access maps (Figures 13 & 14) visually present one of the most common indicators of walkability for urban neighborhoods. Although the distance from each building was not measured through the street network but simply calculated based on a radius, the map showing access to public transportation seats provides an idea of how democratically distributed the transportation facilities are within the municipal boundaries of Kadıköy. Access to greenspace that we created a map for (Figure 15) is a similarly significant contributor to walkability and quality of life. Our exercise routes map (Figure 16) gives an idea of which residential locations have better access to recreational activities within the city. Besides distance, we argue that this map may reveal additional contributors to the exercise behavior in the city such as the pleasantness, comfort and safety of these routes, as well as their accessibility from different neighborhoods. Municipalities and central authorities can manage the allocation of their resources based on information regarding the distribution of public services and facilities easily observable through our maps demonstrating access to public transport, green areas and routes used for sports activity. Since residential neighborhoods less fortunate in terms of access to public transportation can quickly be identified in these maps, they can encourage local residents to demand an increase of routes and more frequent services from transportation authorities. Access to green areas and the shoreline

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

which is obviously attractive for exercisers can be enhanced by new bike lanes and improved street conditions for better walkability especially in areas that link the northern parts of Kadıköy to these areas. We present these observations and suggestions for urban designers through the case study of Kadıköy, however, we aim this study and these maps to constitute a model for revealing and communicating information that can guide urban decisions in improving the street life, commercial activity, local development and life quality of urban residents through better allocation and more democratic geographic distribution of resources. Without doubt, each urban area from each country is unique in many aspects that also contribute to the urban qualities that we propose can be improved through our observations, thus all data should be mapped and interpreted with various such differences in mind to be relevant in supporting urban decisions. 5. Discussion We do not present any geostatistical analysis of urban data in this paper, but rather propose this research as a preliminary step to be followed by rigorous spatial analyses in studies to follow. Nevertheless, we propose the methodology we present for urban data collection and visualization as a guide for designers to better understand and utilize means to gather and process data relevant for urban design decision support. We acknowledge that there may be possible errors in information provided by the listing owners regarding the real-estate data we utilize. Voluntarily generated data may also cause inaccuracies due to the free data-entry format, as in the case for social media entries. The social media users tend to create their own use of language, abbreviations, and personal methods of using hashtags (Schreck & Keim, 2013) or simply make misspellings. Furthermore, Turkish is an agglutinative language and multiple suffixes make it unfavorable for text based analytics. Additionally, the sources we have used for this research are limited: There are several websites and applications dedicated for similar purpos-

es, and at this stage of the study, the datasets were created using the most widely used website and/or application in their respective category. While this approach gave us a strong initial start, additional sources may be added to the datasets in further research. In such a case, a mutual database format should be created as the type of data retrieved from different sources will also begin to differ. Final discussion marks include copyright issues; as almost all of the websites used in these studies limit reproduction and re-distribution of the data for commercial use in their legal user agreements. Boeing and Waddell (2016) note a case where a federal US court decided that web scraping publicly available data did not constitute a copyright violation. Even though non-commercial research does not fall under legal restrictions, only a handful of the websites and applications make the data collecting process openly available to developers and researches. This means that the sources without developer access may easily limit or restrict our access in the future by making changes to their web structures. 6. Conclusion Nowadays, a significant portion of public and commercial activity takes place on online social networks. We swim in an enormous pool of data streaming all around us. The tweets, selfies, check-ins, likes, ignores, pokes, snaps, listings, stars, user comments constantly keep updating and adding to this pool. Mostly consisting of unorganized, unfiltered and uncategorized raw data; this pool is the ultimate collective subconscious of our society. It has the information on what people like doing where, when, how; how they travel; how they spend their time; how they speak to each other; what their share of economy is; how they take what was designed and how they are transformed by that. This is a valuable source of information for anyone interested in exploring the geospatial past, current time and future in great detail. Although urban designers and authorities already make use of the data available to them, traditional data collecting methods tend to be more static

Web scraping and mapping urban data to support urban design decisions

and data is not so easy to update, whereas spatial information is exponentially more valuable when the time layer is included in the equation. Deriving the data from constantly self-updating sources, and establishing a workflow that collects, organizes, and processes this data ensures access to up to date scenarios at all times. Additionally, the accumulation of this data creates a spatio-temporal archive in time, which would illustrate the development and evolution of urban use patterns. Through this research we have integrated different computational approaches together to create an automated workflow that draws data from online sources and visualizes them on maps. A major part of this workflow can be utilized to draw data from various other sources with minor tweaks in the query code. We present multiple layers of urban data that we believe to be relevant for aiding urban design decisions by revealing various urban use patterns and distribution of public and commercial amenities and services within a neighborhood. We visualize our data layers on maps which we propose to constitute a basis for further study where geospatial statistical analyses will be performed to achieve quantitative results. We believe that with the further development of web technologies and integration of mobile devices into our lives, the definition of a city and what it encompasses is already being redefined fundamentally. Therefore, approaches to urban analysis and design must be re-evaluated and become as dynamic our urban environments already have. Acknowledgements The authors would like to thank Can Kadir Sucuoğlu and Cemal Koray Bingöl for their outstanding contributions in this study. References Balaban, O., & Tuncer, B. (2016). Visualizing Urban Sports Movement. In Complexity & Simplicity - Proceedings of the 34th eCAADe Conference (Vol. 2, pp. 89–94). Oululu, Finland: eCAADe and University of Oulu. Batty, M. (2013). Urban Informatics and Big Data. London.

Batty, M., Axhausen, K., Giannotti, F., Pozdnoukhov, A., Bazzani, A., Wachowicz, M., Ouzounis, G., Portugali, Y. (2012). Smart cities of the future. European Physical Journal: Special Topics, 214, 481–518. Bency, A., Rallapalli, S., Ganti, R., Srivatsa, M., & Manjunath, B. (2017). Beyond Spatial Auto-Regressive Models: Predicting Housing Prices with Satellite Imagery. In 2017 IEEE Winter Conference on Applications of Computer Vision (pp. 320–329). Santa Rosa, CA, USA: IEEE. Boeing, G., & Waddell, P. (2016). New Insights into Rental Housing Markets across the United States: Web Scraping and Analyzing Craigslist Rental Listings. Journal of Planning Education and Research, 1–20. Chen, N. C. (2012). Urban Data Mining: Social Media Data Analysis as a Complementary Tool for Urban Design. MIT. Clarke, A., & Steele, R. (2011). How personal fitness data can be re-used by smart cities. In 7th International Conference on Intelligent Sensors, Sensor Networks and Information Processing (pp. 395–400). Adelaide, SA, Australia: IEEE. Cohen, J. P., & Coughlin, C. C. (2008). Spatial hedonic models of airport noise, proximity, and housing prices. Journal of Regional Science, 48(5), 859–878. Cranshaw, J., Schwartz, R., Hong, J. I., & Sadeh, N. (2012). The Livehoods Project: Utilizing Social Media to Understand the Dynamics of a City. In Proceedings of the Sixth International AAAI Conference on Weblogs and Social Media (pp. 58–65). Palo Alto, CA, USA: The AAAI Press. Geoghegan, J., Wainger, L. a., & Bockstael, N. E. (1997). Analysis Spatial landscape indices in a hedonic framework:an ecological economics analysis using GIS. Ecological Economics, 23, 251–264. Glaeser, E. L., Duke-Kominers, S., Luca, M., & Nalk, N. (2015). Big Data and Big Cities: The Promises and Limitations of Improved Measures of Urban Life. HKS Faculty Research Working Paper Series. Cambridge, MA. Jacobs, J. (1961). The Death and Life of Great American Cities. New York,

ITU A|Z • Vol 15 No 1 • March 2018 • E. Ensari, B. Kobaş

NY, USA: Random House. Jenkins, A., Croitoru, A., Crooks, A. T., & Stefanidis, A. (2016). Crowdsourcing a collective sense of place. PLoS ONE, 11(4), 1–20. Jiang, S., Alves, A., Rodrigues, F., Ferreira, J., & Pereira, F. C. (2015). Mining point-of-interest data from social networks for urban land use classification and disaggregation. Computers, Environment and Urban Systems, 53, 36–46. Kurgan, L. (2013). Million-Dollar Blocks. In Close up at a distance: mapping, technology, and politics. Zone Books. Lee, J.-G., & Kang, M. (2015). Geospatial Big Data: Challenges and Opportunities. Big Data Research, 2(2), 74–81. Li, S., Ye, X., Lee, J., Gong, J., & Qin, C. (2017). Spatiotemporal Analysis of Housing Prices in China: A Big Data Perspective. Applied Spatial Analysis and Policy, 10(3), 421–433. Niederer, S., Colombo, G., Mauri, M., & Azzi, M. (2015). Street-level City Analytics: Mapping the Amsterdam Knowledge Mile. In Hybrid City 2015: Data to the People (pp. 215–220). URIAC.

Pettit, C., Widjaja, I., Russo, P., Sinnott, R., Stimson, R., & Tomko, M. (2012). Visualisation support for exploring urban space and place. ISPRS Annals of Photogrammetry, Remote Sensing and Spatial Information Sciences, I-2, 153–158. Schreck, T., & Keim, D. (2013). Visual Analysis of Social Media Data. Computer, 46(5), 68–75. Peng, D., Biagi, L., & Ito, U. (2016). Leading countries based on number of monthly active Instagram users as of 1st quarter 2016. Retrieved October 5, 2017, from http://www.statista.com Townsend, A. M. (2013). Smart Cities: Big Data, Civic Hackers, and the Quest for a New Utopia. New York: W. W. Norton & Company Inc. Waddell, P., Berry, B. J. L., & Hoch, I. (1993). Residential property values in a multinodal urban area: New evidence on the implicit price of location. The Journal of Real Estate Finance and Economics, 7(2), 117–141. Yemeksepeti’nden 2016 lezzet almanağı. (2016). Retrieved October 5, 2017, from http://www.yemeksepeti. com/

Web scraping and mapping urban data to support urban design decisions

ITU A|Z • Vol 15 No 1 • March 2018 • 23-38

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

Pınar ÇALIŞIR ADEM1, Gülen ÇAĞDAŞ2 1 pinarrcalisir@gmail.com • Department of Architecture, Graduate School of Science Engineering and Technology, Architectural Design Computing Graduate Program, Istanbul Technical University, Istanbul, Turkey 2 glcagdas@gmail.com • Department of Architecture, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.40316

Received: October 2017 • Final Acceptance: November 2017

Abstract Traditional cities are complex organic systems which have many forms and structures affecting each other spontaneously through time. Today, uncontrolled urbanization activities destroy the wholeness of these organic patterned structures and destroy their local character by top-down planning decisions and designers’ personal motives. Thus, for enhancing the continuity in traditional urban patterns, understanding the inner nature of urban patterns is crucial. In this context, the aim of the proposed method is to deal with the huge number of raw data coming from complex traditional cities and analyze them with computational techniques in order to let designers create some basic rules and understanding in terms of the spatial organization of city structures in the early phase of urban design process. These rules can give clues about the essence of the city and guide designers and authorities for better integrated designs with traditional patterns not only in their physical form but also in the social, economic and cultural context. In this sense, the proposed method offers Data Mining algorithms in terms of knowledge discovery in the urban database with the help of Geographic Information System (GIS) technologies. Keywords Data mining, Urban patterns, Traditional city.

1. Introduction Recently, modern urban planning approaches have deteriorated the relationship between the newly designed area and the context, which is embodied in traditions and represents a sense of place, as a result of handling the urban structure in parts by considering them unrelated. Besides, top-down planning concepts and schemes have controlled self-organized, unpredictable and slow growth of traditional cities. Aforementioned applications caused cities to become standardized and lose their true essence which gives them their identities. Since profit-base planning decisions cause uncontrolled urbanization, discontinuity in the traditional city structure remains today. Urbanization activities, such as high-density housing and standardized apartments, oversimplify the traditional urban form and create a built environment which is alien and disconnects with local culture, history and life rituals. Thus, the years of natural spatial evolution of traditional cities and the healthy functioning of these environments are damaged. However, the traditional city is a complex system like a living organism which has many forms and structures affecting each other spontaneously. Even if, we shape specific areas of the traditional city with complete and rigid visions of laws and planning decisions, through time, it reorganizes its parts and transforms with bottom-up forces. Therefore, decomposing and analyzing parts of the city and their relationships are significant to reveal the inner nature and different layers of this urban system. Only with this understanding, we can protect the natural evolution of traditional city patterns and design new harmonious settlements. Collecting as much as data from the city and making sense of them are crucial for making further decision in new design processes. Natural continuity between new and old settlements can be maintained provided that the collected data turned into useful knowledge about existing structures and following periods built on this knowledge. On the other hand, the built environment supplies designers with vast volumes of data to analyze and interpret.

Therefore, most of the time, analysis of urban structures and discovering relationships among parts of the traditional city become superficial, inadequate and depends on designer’s personal motives. Furthermore, repetitive patterns, random behaviors and different urban attributes influencing one another in the structure of the city are ignored by designers. So, interferences happen in the natural spatial evolution of traditional cities, and new settlements arise as imitations of old ones or become incompatible with the existing city form. In this context, this study aims to propose a method to objectively evaluate and analyse a vast number of raw data which are accumulated beneath patterns of complex traditional cities, with computational techniques to let designers develop an understanding regarding the spatial organization of city structures. This knowledge can be turned into basic rules to be used in the early phase of urban design processes, providing designers and authorities with clues about the essence of the city for better-integrated designs with traditional patterns in their physical form. The method can discover the hidden relationships in the structure of the traditional city and let designers to enhance the continuity in the traditional urban patterns by avoiding subjective interpretations and implementations of designers and authorities in the traditional patterns. In this sense, the proposed method offers Data Mining algorithms for automated and computational discovery of knowledge in the urban database with the help of Geographic Information System (GIS) technologies. 2. Complex nature of traditional cities Traditional cities are complex entities with organic patterns evolved in many years with various determinants. Traditional cities have not only “organic patterns” (Kostof, 1991, p. 43) but also behave like “growing organisms” (Alexander et al., 1987, p. 13) creating complexity. The term “the city with organic patterns” expresses that the city comes into being after a long process without the master plan or the assistance of designers, but takes shape

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

with daily routines of inhabitants and the topography. According to Kostof (1991), organic patterned cities are shaped by local builders, and they are “spontaneous, chance-grown, generated and geomorphic” (Kostof, 1991, p. 43). Narrow streets compatible with the topography, buildings shared similar typologies, and open spaces dispersed randomly around the city throughout the spatial evolution process (Kostof, 1991). On the other hand, Alexander et al. (1987) expressed that traditional cities have always protected their core structure and advance naturally as a whole like “growing organisms” (Alexander et al., 1987, p. 13). Integrity of the city structure is the actual determinant of its future form and capacity for the city growth is unforeseeable. Based on these theories can be asserted that traditional cities are self-organizing growing systems, starting from a particular core structure, such as living organisms. Growth process gradually takes place and the borders of the growth are unpredictable, whereas it is always harmonious with the general city pattern, because it is primarily fed by existing urban structure. The term ‘city as an organism’ in this study is more than an abstraction for an analogy of biological form. It is related to traditional surroundings and the reasons of their complex character at the end of their spatial evolution and growth through time under the forces of the natural environment and purposes of their inhabitants. In this context, traditional cities can be seen as complex organizations of parts to create integrity which is more than the sum of its parts (Carmona et al., 2003). Unitary visions of city models, which produce “artificial cities with a hierarchy like a tree” (Alexander, 1965, p. 58), emphasize on the overall appearance of cities and ignore working mechanisms of city components which create the actual complexity. In “natural cities” (Alexander, 1965, p. 58), which have organic patterns, global form and behavior have been generated for many years by local interactions between city parts with self-similar characteristics through a network which is “a semi-lattice” (Alexander, 1965, p. 58). Same with living organisms, during the time,

parts of the traditional city cannot comply with the environment disappear or to be forced to change to keep alive. Thus, the evolution of traditional cities through time is the main similarity with living organisms and it maintains natural growth of city patterns based on the core structure of the city. In this process, local masters who shape the city and specialized in traditional techniques create a dynamic/vital structure which can evolve by changing, transforming and adapting itself. Proposed study sees that traditional cities are organic not just for their organic patterns congruent with the natural environment, but also for their complex nature that is a shared feature of living organisms. Therefore, traditional cities are defined as living organisms and treated as complex systems. Also, the proposed method is designed for understanding the self-organized formative process of the traditional city, shaped by numerous determinants. Thus, in urban design activities, especially in terms of new settlement designs, it is necessary to understand these determinants. In order to achieve that the design area and its surrounding need to be analyzed by both qualitative and quantitative research methods. Later, designers can clearly see repetitive patterns, random behaviors and different factors affecting one another in the structure of the traditional city. However, traditional techniques are not capable of that type of evaluation and interpretation of traditional cities’ complex nature. Therefore, we have to propose new methods which are more sensitive in terms dealing with the complex nature of these cities. In this context, we can find various studies regarding to complexity of traditional towns. For instance, Gürbüz et al. (2010) propose fractals to understand the city structures focusing on their geometrical and dimensional features and offer new design alternatives enhancing the continuity of traditional urban patterns. Similarly; Duarte et al. (2006), utilizes shape grammars by considering functional and morphological features of urban structures for preserving the character of traditional urban structures while proposing new design areas to fulfill modern life

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

requirements. Karabağ (2010), in his Ph.D. thesis, offers a methodology for analyzing traditional urban patterns and their relations regarding their architectural features via statistical tools and generate rule sets for an algorithmic design procedure for new design interventions with the results coming from the data analysis phase. In this article, Data Mining techniques are employed to handle complexity in traditional settings. The method is proposed for analyzing urban data, which can be both categorical and nominal, finding their relationships and constituting an understanding of the organization of traditional urban character through basic formation rules. As we mentioned at the beginning of this chapter; the traditional city produces a massive amount of data based on its complex organic nature and extracting useful information from this complex structure is unfeasible with manual analysis. However, Data Mining can handle a huge amount of raw data in a concise time without the years of training as an analyst and contains various algorithms for analyzing the raw urban data in very different perspectives, such as clustering, frequent pattern mining, classification and so on. Therefore, in this study, we choose to utilize from Data Mining algorithms for knowledge discoveries in the complex traditional urban patterns. Enhancing different Data Mining techniques cannot only help us comprehend urban attributes and their relations from different perspectives but also find anomalies in the traditional patterns. In this way, we can protect city’s self-evolved and chancegrown urban structure while new settlements are designed by considering repetitive and unique structures. 3. A method proposal for interpretation of traditional urban data by using data mining techniques For harmonious new settlements and the protection of traditional city’s organic essence, designers must have definite ideas about architecture, social, cultural and economic dynamics of the city in addition to its history and local values. Therefore, the design area and surroundings need to be analysed

in depth. Results coming from these analyses may help designers to reveal structures of traditional cities. In order to do this, we have to collect as much as information from cities and find appropriate techniques which can resolve the complex structure of traditional urban systems. In this context, the proposed method utilizes Data Mining algorithms. Data Mining is a process for finding latent connections and extracting novel and logical inferences from raw data in huge databases (Hand et al., 2001). In other words, “Data Mining is defined as the identification of interesting structure in data” (Fayyad et al., 2002, p. 28). It belongs to a larger context called Knowledge Discovery in Databases (KDD) which “makes sense of the data” (Fayyad et al., 1996, p. 37). Every day, heavy load of information is uploaded into databases, and this raw data cannot be analysed manually. KDD, which arose during the 1980s (Sohtorik, 2016), provides us with computational tools to examine, interpret this data and construct plausible hypotheses for our interest in an automated and objective way. This process of transformation from raw data to useful knowledge allows us to analyse the current situations, make predictions and decisions for the future. Data Mining is the main part of this process which is “the application of specific algorithms for extracting patterns from data” (Fayyad et al., 1996, p. 39). Patterns are specific to the domain from where data are collected. Based on the intentions of the analyst, patterns are evaluated and considered as knowledge according to their “validity, novelty, usefulness and simplicity” (Fayyad et al., 1996, p. 41). In this context, Data Mining contains different mathematical techniques for various tasks, producing patterns from transformed data in databases for further interpretation and evaluation. These tasks can be classification, segmentation and clustering, association, deviations, trends and regression analysis and generalizations (Miller and Han, 2009). Thus, in Data Mining studies, it is crucial to choose appropriate algorithms for the data type, data scale and the aim of the analyst in order to obtain meaningful and reliable results. Looking for patterns in

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

Figure 1. The phases of the proposed method.

databases is not unique to Data Mining. Different fields -such as statistics, pattern recognition, and exploratory data analysis- apply various techniques for handling this issue. Data Mining borrows different algorithms from these areas and offers semi or fully automated, easy to use, accessible and practical tools to evaluate large databases and discover useful patterns for users without a significant training as data analysts (Fayyad et al., 2002). Various studies are utilizing Data Mining techniques in the architectural context. For example, by using Data Mining techniques, Gill et al. (2009) define urban typologies in the street and block level. Reffat (2008), searches patterns in features of contemporary Arabic architecture. Sohtorik et al. (2010), investigate patterns and relationships of urban attributes and propose a Data Mining methodology for new urban interventions in the complex nature of the city (Sohtorik, 2016). Laskari et al. (2008), quantify spatial attributes derived from plan features and finding related patterns of different spaces in the scale of urban blocks. Hanna (2007), defines archetypes from building examples and utilizes them into new design processes. laskari The proposed method in this article has three consecutive phases as illustrated in Figure 1. The phases are: • The Formation of Data • The Analyses of Data • The Interpretation of Data Throughout the proposed method, to compile and visualize different sources of data from the traditional city GIS software-ArcMap/ESRI was

used. Cleaning digital data was done in AutoCAD/Autodesk. In the analysis phase of this data, RapidMiner –opensource software- was used. In the first phase of the proposed method, which is called “the formation of data”, a database for attributes in the traditional city is created. Attributes, which carry information, are features of urban entities such as urban blocks, buildings, and streets. The urban database should be built in order to collect and store various features of urban entities and include as much as authentic and official data –both quantitative and qualitative- about the existing setting. Therefore, sampling of raw data should be done using various sources such as municipalities and different government agencies. Furthermore, if there is a need for additional information about the urban structure, designers can generate “An Info Form” to increase the data. In this study, GIS software-ArcMap is used for gathering raw data coming from different sources and creating the database as an Attribute table. Also, spatial analyst tools inside ArcMap generate additional information for slope, distance and aspect attributes. Through the database, collected data becomes unique to the neighborhood. From Table 1, we can see that titles in the attribute table which may vary in a wide range of data; such as architectural features, building location, dimensional features, land use information, topography and so forth. Also, data in the attribute table can be classified into different groups. In the present context, we have numerical values and categorical values which are called nominal

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

Table 1. Attribute table for the traditional urban database.

based on their non-hierarchical structure. Throughout the study, additional attributes can be added or unnecessary data can be eliminated from the table in line with the selection of designers. The second phase of the method includes Data Mining techniques and is called “the analyses of data”. It is the primary phase of the method in which analysts or designers aim to analyse attributes by Data Mining algorithms in the database concerning their latent connections, repeated patterns or unique formations to reveal the complex nature of the traditional surroundings. If we consider the complex traditional city as a giant database accumulating raw data beneath its spatial organizations, we can use Data Mining techniques to produce useful knowledge by collecting, selecting and evaluating data focusing on our design problems. These methods can help us easily understand the organizational characteristics of urban entities and different algorithms may help us to investigate raw data from the very different point of views in an automated and objective way. In this way, we can detect unfamiliar or dominant patterns, relationships among city elements and interpret these findings to reveal the essence of the traditional city which gives form to it.

In this scope, we applied two different Data Mining algorithms. First one is clustering algorithm which is used to detect subsets of the target data. Clustering is a technique for sparing a dataset into subsets or clusters based on their common features. The primary goal of the clustering is to collect similar objects in the same cluster and put different objects in the separate clusters as much as possible (Han et al., 2011). In the proposed method, the clustering algorithm can help us reveal subsets of the database which contains buildings sharing similar spatial features. Clustering techniques are highly used in image pattern recognition studies, web search techniques, fraud detection and biology disciplines. There are various clustering algorithms according to data scale and data type. In this study, the mixed nature of our dataset and the difficulty of predicting cluster numbers beforehand lead us to use DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. Shapes of data points inside the database are not regular all the time. Sometimes, these shapes can cumulate in specific places and create arbitrary shapes in the database. DBSCAN algorithm works with large databases, especially with random shapes such as in spatial databases (Ester et al., 1996).

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

Figure 2. DBSCAN clustering algorithm process in RapidMiner.

Figure 3. An example of an association rule and how it is defined by the support and the confidence values (Han et al., 2011).

Figure 4. Frequent pattern mining process in RapidMiner.

DBSCAN algorithm determines the local regions on which these points concentrate, rather than merely measuring the distances between points (Zaki and Meira Jr., 2014). Overall, the primary strategy of DBSCAN algorithm is to identify clusters according to dense regions of data points which are irregular, such as ‘S’ or ‘oval’ shape (Han et al., 2011). Also, this algorithm needs only two user-specified parameters: Epsilon and MinPts. Epsilon determines the neighborhood radius, and MinPts defines the minimum point of numbers inside the neighborhood. Based on these parameters, density is calculated by counting the number of data points inside the neighborhood specified by Epsilon value. DBSCAN algorithm is easy to use with RapidMiner interface. Also, it can be applied both numerical and categorical data as well as the mixed nature of data due to Mixed Euclidean Distance measure type in RapidMiner. From Figure 2, we can see the process of the DBSCAN model inside the RapidMiner software. The second algorithm used in this study is Frequent Pattern-Growth (FPGrowth) Algorithm. This algorithm finds frequent patterns in the dataset and produces association rules for urban attributes. The algorithm belongs to a broader term - Frequent Pattern Mining which contains different algorithms to find frequent subsets in the given database. Frequent Pattern Min-

ing algorithms are constructive in revealing useful and meaningful patterns in the dataset. The main objective of these algorithms is displaying relationships between objects and their attributes in the database by finding hidden trends and behaviors (Zaki and Meira Jr., 2014). From the results coming from algorithms, we can define Association Rules for target objects and their attributes. In order to do that, two values need to be considered: the first one is the Support value and the second one is the Confidence value. The evaluation of the results and Association Rules can be achieved by considering these two values. From Figure 3, we can see an example of an Association Rule and how it is defined by the Support and the Confidence values (Figure 3). Besides, FP-Growth Algorithm finds frequent patterns in the dataset and produces association rules for urban attributes and works on an item set by dividing and editing its elements according to a frequency value. Frequency value determines how often an element occurs in an itemset. The algorithm creates a tree structure in order to keep track of subsets, and by doing so, it prevents repeating objects from being held in the memory. In Figure 4, Association Rules model in RapidMiner is illustrated as consecutive steps. In the last phase of the proposed method, which is called “the interpretation of data”, designer or analyst

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

Figure 5. Hatuniye Neighborhood: Monumental areas and traditional buildings.

should interpret data mining results and reveal rules or relationships hidden under the traditional urban pattern. These rules or patterns represent the essence of the traditional city and reflect its characteristics. Since rules come from an objective assessment of urban attributes and their relationships, a rationale can be constituted regarding new settlement designs to enhance the continuity of the pattern of the traditional city. 4. Case study: The traditional city of Amasya and Hatuniye Neighborhood For the implementation of proposed method, historic Hatuniye Neighborhood in Amasya/Turkey was chosen and ArcMap was utilized to create a database for this neighborhood. Amasya is one of the important historical cities in Turkey, which dates back to the 3000 BC according to archaeological excavations in the city center (Özdemir, 1996). The city develops linearly in the valley opened by the Iris River (Yeşilırmak) and is surrounded by sharp mountains on all four sides. Since the Chalcolithic Age; Pontus, Byzantine, Seljuk, and Ottoman dominance reigned over the city. Therefore, the urban configuration of the city is formed as multi-layered. In many quarters of Amasya, current patterns of streets and buildings melt into patterns of the old city so that one cannot be altered without the other. Today, we can witness the historic urban pattern from the Ottoman Empire and houses date back to 300 years

ago. But like many traditional settings, the center of the city is under threat of contemporary urbanization which is not sensitive to local climate, topography, and culture and has no respect for the continuity of the traditional pattern as seen in Figure 6. Thus, there is an urgent need for diversifying computational methods, which can handle the complexity of this organic city part, to understand the real essence and organizational characteristics of the traditional city in order to protect the city’s self-evolved structure respecting local climate, topography, and culture. Hatuniye Neighborhood, which is one of the historic Neighborhood of Amasya, also called “Inner City” is located along the Iris river and leans its back to the Kırklar Mountain. At the peak of the mountain, Harşena castle, above it 5 Pontic tombs and the urban structure of the neighborhood with the river create “a poetic urban experience” (Bechhoeffer and Yalçın 1991, p. 24). Despite massive destructions throughout the history, this quarter protects itself thanks to its unique location and surrounding city walls, which create barriers with river and mountain position and separate the neighborhood from the rest of the city. From Figure 5, we can see the general characteristic of the neighborhood. The neighborhood has four bridges, and two of them draw the periphery of the neighborhood. For a general view, the quarter consists of 14 street blocks, 204 parcels, 165 main buildings and 41

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

Figure 6. Amasya view with Hatuniye Quarter: Confrontation of the traditional and contemporary pattern.

Figure 7. Closer view to Hatuniye Quarter from Iris River: City walls as foundations for buildings.

Figure 8. Building examples specific to the neighborhood: The building on the left is “Selamlıklı” (Redrawn from Amasya İli, Hatuniye Mahallesi Geleneksel Yerleşim Dokusunun Analizi, Değerlendirilmesi Ve Koruma Geliştirme Önerisi (Unpublished Master Thesis), by Türkoğlu, E.,2006, Ankara: Gazi University; Institute of Science).

additional buildings in total. Most of the waterfront houses in the Neighborhood are from the Ottomans in the 19th century. In general, houses were built either in the “bağdadi” or “hımış” techniques. Nevertheless, foundations of waterfront houses were made from cut-stone (Yalçın, 1998). These stone foundations are part of the city walls and riverfront buildings located upon them (Figure 7). In other words, in the ancient times, the city wall was merely a protection for the city. Later, the Ottomans used traces of ancient neighborhood and built their houses upon them. Therefore the city wall did not served as a protection, but a masonry wall for their foundation. We can trace the Ottoman city patterns from the general view of the neighborhood and these patterns reflect great complexity in self-similar, self-orga-

nized, and dynamic urban structure which is constantly and slowly change like a living organism. The floor plans in Figure 8 illustrate connections which buildings establish with their neighbors and the public space. Despite huge construction activities, that ruined most of the traditional buildings in the quarter, there are still traces of Ottoman urban structure through the streets. Therefore, implementation of the proposed method in this quarter may help local designers and authorities to gain a deeper understanding of this unique structure and its underlying principles for future interventions. 5. Implementation of the proposed method in Hatuniye Neighborhood As we mentioned in Chapter 3, the proposed method subdivided into

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

three steps: “the formation of data”; “the analyses of data” and “the interpretation of data”. In the formation phase of the method, we created a database through ArcMap, preprocess and visualize the data for Data Mining algorithms. Firstly, we gathered city information through Amasya Municipality in the format of vector maps. These AutoCAD maps contain official urban data about the neighborhood. Cleaning and selecting useful layers inside maps were done in AutoCAD. However, the collected information from this institution was not enough for more in-depth and reliable operations; therefore, creating a building info form to increase the data was necessary. As illustrated in Chapter 4, floor plans of buildings were collected from Türkoğlu’s master thesis (2006) to gain a better understanding of structures in the context of the neighborhood setting. The info form about buildings holds various data which are values of attributes of urban entities previously represented in Table 1. After collecting as much as data, we started visualizing the urban data in ArcMap and matched the data with actual buildings. ArcMap does not only visualize the data in the real space, but also it developed our data with its spatial analyst tools such as slope and aspect operations. Moreover, with Multiple Buffer Ring tool, we classified all buildings, according to their distance to monumental buildings. ArcMap stored all the urban data matched with buildings in the Attribute Table, and we could export this table in Excel format to use in the analysis process with Data Mining. In the analysis phase of the method, the data table from ArcMap was imported in a RapidMiner as an Excel sheet. For a preliminary action, software analysed all nominal and numeric values, such as parcel and building areas, in terms of maximum, minimum and average values. Based on these analyses, the smallest street block is 66,468 m² and the biggest one is 9144 m². For the size of parcels, while the smallest one is 21,196 m², the biggest one is 949,392 m², the average value of parcels is 156,416 m² and deviation value is 118.856. Similarly, the smallest building size is 21,196 m² while the

biggest one is 324,786m² with average 83,031 m² and the deviation value 53.180. Monumental buildings with large areas were also added into the calculation as we can understand from high value of deviation, therefore, their influence on the average should be considered during the interpretation process. In this way, Rapid Miner can give us statistical results about maximum, minimum and average values for the building envelopes and open spaces for further design studies in order to capture the scale of the existing traditional pattern and avoid discontinuities in the visual integrity of the neighborhood. Also, in terms of nominal values, 146 buildings out of 165 is in the low-slope area (slope value: 0-24%); while only 19 buildings in the high-slope area (24-63%). 137 buildings out of 165 are in the south aspect, while 12 buildings in the southeast and 16 buildings in the west aspect. Only 5 buildings have oriel windows and 15 buildings out of 165 have “selamlık” building (which is an additional room -or building in Amasya case- for male guests in the Ottoman plan layout). 98 buildings do not have any projection on their façade and 89 buildings do not have a hall. 98 buildings out of 165 do not have basement floor and 42 buildings do not have a courtyard. 61 buildings are above ground level with high basement wall. 108 buildings are attached house and 88 buildings have a main entrance from the street to the courtyard, while 53 buildings have direct access to the house from the street. These statistical results can give an idea about the general characteristics of the neighborhood. However, for gaining deeper insight about urban attributes and their relationships, more Data Mining algorithms were applied. In the data table for attributes of Hatuniye neighborhood, there are mainly nominal but also few numerical attributes; so we cannot foresee the number of clusters beforehand. Therefore, DBSCAN algorithm was used to see consistent patterns or anomalies in the urban structure. As we mentioned in Chapter 3, DBSCAN operations need two user-specified parameters: Epsilon and MinPts. In RapidMiner default value for Epsilon is 1.0 when the MinPts val-

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

Table 2. DBSCAN clustering algorithm. Clusters of Test 1.

ue is 5. We mostly followed MinPts value, unless we did not need more refined or crowded clusters. Low or High number of epsilon value did not give us any feedback; therefore, we set this number to 1.0 in all DBSCAN operations. Additionally, we chose our distance measure, according to our attribute types (the nominal, numerical or mixed type of attributes). First of all, we chose all attributes for DBSCAN algorithm with chosen parameters. But, the model cannot produce any cluster. In order to gain more refined clustering results, we choose some attribute groups which repeat in the neighborhood from a general point of view. Thus, in test 1, we set Epsilon to 1, and MinPts to 5, for revealing attribute relationship between “additional building existence” and “distance to the neighborhood square”. Results after of the clustering process can be seen in Table 2 as an example of clustering operations. Test 1 gave us 11 clusters with Cluster 0 containing six buildings as noise data. According to the results, we can specify that 56 buildings which are 0-200 meters away from the neighborhood square do not have any additional building and the probability of having additional structures increases as buildings become distant from the neighborhood square. In test 2, we wanted to discover the relationship between “additional building existence” and “parcel area” with parameters of

test 1. The algorithm created 11 clusters with 19 noise data inside Cluster 0. According to results, there are 45 buildings with no additional structures and their parcel base is between 0-100 m². Buildings in the 100-250 m² parcels are at similar numbers. In test 3, chosen attributes are “view” and “basement existence” with the same parameters with previous tests. At the end of the algorithm, there were 7 clusters with 1 noise data in Cluster 0. The most important clusters at the end of the operation are Cluster 1 and Cluster 2. Cluster 1 contains 56 buildings with street view and no basement floor. Cluster 2 contains 52 buildings with river view and basement floor. Cluster 3 contains 26 buildings with no basement floor in the river area, and there are 8 buildings with street view and the basement floor in Cluster 4. Cluster 5 contains 6 buildings with basement and Cluster 6 contains 16 buildings with no basement and view of the mountain. As seen from results, in the riverside, there are more buildings with basement floor and buildings with the view of the mountain or street usually do not have a basement floor. In test 4, selected attributes were “view” and “courtyard location”. The most important cluster, in this case, is Cluster 2 containing 62 buildings with river view and a front courtyard. Also in cluster 4, there are 10 buildings with an inner courtyard and river view. Only 5 buildings on the riverside do not have a courtyard which is represented in Cluster 3. In test 5, we considered “aspect” and “building position” attributes. According to the results, there were 6 clusters, including Cluster 0 with 8 noise data. In this test, the most important cluster is Cluster 3 with 84 attached houses in the southern aspect. When “building position” was compared to “view” attribute in test 6, results appear in a similar way. Based on the results, in Cluster 3, which is the most crowded one, there are 71 attached houses on the riverside. In test 7, “aspect” and “courtyard” locations were considered. As a result, 7 clusters appeared with 17 noise data in Cluster 0. The most important cluster is Cluster 2 with 77 buildings in the south aspect with a front courtyard. In test 8, we wanted to discover “hall-

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

type” and “parcel area” relationship. But when we used the same MinPts value with the previous test, which is 5, we get 13 clusters including Cluster 0 with 57 noise data. Therefore, in test 8 MinPts was set to 3. According to the results, 30 clusters appeared with 23 noise data in Cluster 0. Here, buildings with no hall were detected. Cluster 1 appears with 11 buildings with no hall and 0-50 m² parcel area. Cluster 16 has 29 buildings with no hall and 50-100 m² parcel area. Cluster 3 has 18 buildings with no hall and 100-150 m² parcel area. Cluster 6 contains 10 buildings with no hall and 150-200 m² parcel area. Cluster 9 and 10 has 6 buildings with no hall and 200-250 m² and 250-300 m² parcel area. Cluster 11 has 2 buildings with no hall and 350400 m² parcel area. Finally, Cluster 24 and Cluster 29 have 2 buildings with no hall and 300-350 and 500-550 m² parcel area. As a result of the clustering studies, we can assert that clustering techniques can help us to see repetitive patterns, anomalies or obvious structures in the urban pattern. Another Data Mining technique used in this study is the Frequent Pattern Mining which reveals attributes that are frequently used together and create some association rules according to mining results. FP-Growth algorithm is used in these tests as we mentioned in Chapter 3. The results of the first test can be seen in Table 3. In the first test, we tried to use all nominal attributes for the creation of association rules. By changing support and confidence values we may create a high number of Association Rules. In this case, our support value is 0.5 and confidence value is changing between 0.7 to 0.9. In Table 3, we can see association rules which define urban attributes frequently used together in the dataset. The first column represents premises and the second column represents conclusions. Other columns give us confidence and support values according to attributes relationship between the first two columns. For instance, according to rule. 1 values, we can say that 56% of all buildings in the neighborhood have no additional structures and oriel windows in the south aspect and they have an entrance

Table 3. Selected association rules for Hatuniye Neighborhood (The first test).

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

Table 4. The second association rule test.

Table 5. The third association rule test.

from north-south orientation. However, 73% of buildings with no additional structures in the south aspect has an entrance from north-south orientation. For rule no.2, 55% of all buildings in the neighborhood lay on the low slope surface, constructed in the south aspect with no additions and balcony; whereas 73% of buildings constructed in the low slope area and oriented in the south aspect has no additional buildings or balconies. For an over-

all overview of the rules, we can state that 74% of buildings in the south aspect have courtyards. 75% of courtyard buildings are attached houses with no oriel window, and 76% of them have a front courtyard. 89% of buildings with entrances from the ground level is in the south aspect with no additional buildings and entry platform. 90% of buildings in the south aspect and no basement floor are in the low slope area. Also, 90% of buildings with a courtyard in the south aspect has no additional structures. In test 1, although the algorithm produces more association rules and confidence is more than 70%, most of the time the rules are recurring at different attribute combinations. Therefore, we chose Association Rules for their best fit to our objectives. In the following tests, we tried to filter urban attributes for the creation of more focused association rules. For instance, in the second test represented in Table 4, the selected attributes are: “Building Base Area”, “Number of Floors” and “Distance to Governorship”. Confidence value, in this case, was changed to 0.6 and the support value was determined as 0.5. So, the algorithm becomes more sensitive concerning finding hidden relationships between urban attributes. According to this test, more focused rules emerged, as seen in Table 4. From Table 4 we can focus on the rule no. 1 saying that 61% of two-storey buildings 500 meters away from governorship have a base area approximately 50-100 meters square. Also, the rule no. 4 can be useful for revealing an association between “building base area” and “number of floors”. As it turns out, 70% buildings with the base area of 50-100 meters square are two-storey buildings. In test 3 for association rules, the selected attributes were “building base area”, “hall-type” and “parcel area”. This test gave us rules in Table 5. In these rules, we can concentrate on rule 2 which has more meaning than the others. It says that 65% of buildings which have 50 to 100 meters square parcel and the base area has no “hall/sofa” in their plan schemes. Interpretation of these results is the final phase of the proposed method. All

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

buildings in the selected area -Hatuniye Quarter was examined by Data Mining algorithms to determine specific characteristics representing the traditional urban setting. These features can be used in new settlement designs in the neighborhood. Thus, in this section, the results coming from Data Mining operations were interpreted to discuss what this method may offer regarding to investigating urban patterns and defining their relations to each other for urban design studies embracing locality. Significant Data Mining results can be interpreted as follows: Clustering Test 1 and 8: Buildings closer to the neighborhood square usually do not have any additional buildings or hall. Clustering Test 2: If we have parcels around 0 to 100 m² we do not likely design additional buildings due to small parcel areas. Clustering Test 3: In the neighborhood, an essential pattern is buildings with the basement floor and river view. Anomalies in terms of basement floor identify new buildings for accommodation which destroy the integrity of the traditional pattern regarding building scale. Clustering Test 4: In the riverside houses living rooms are usually designed through the river. Therefore the main entrance to the building is on the north and buildings have a front courtyard. Also, only 5 houses do not have a courtyard. So, creating a semi-private area in front of the building is a trend among riverside houses. Clustering Test 5 and 6: Buildings in the south aspect are usually attached houses. The same pattern is also seen in riverside houses. Therefore, new designs in these areas should follow this pattern. Clustering Test 7: Similar to test 5 and 6, in the south aspect, most of the buildings have a front courtyard because of the same reasons with clustering test 4. Living rooms face to the south for better river view in addition to natural light. Assoc. Rules: Test 1. Rule 1: Buildings in the south aspect and without additional buildings are oriented towards North-South. Assoc. Rules: Test 1. Rule 2: Buildings

in the south aspect and low-slope area do not have a balcony or additional buildings. Assoc. Rules: Test 1. Rule 5: Buildings oriented towards North-South have a saddle roof in the same orientation based on the attached house pattern. Assoc. Rules: Test 1. Rule 7-8-9-1011: Buildings with no high foundation walls are usually in the low-slope area and south aspect. Besides, they do not have basement floor. Assoc. Rules: Test 2. Rule 1-4: Buildings away from governorship usually have smaller base areas. Also, if we design houses between the area 50-100 m², we can use 2-storey to follow the urban pattern scale. Assoc. Rules: Test 3. Rule 2: If buildings in small parcels and base areas are designed, using the hall in the plan layouts is unnecessary. These results can be promising for understanding the nature of the neighborhood structure for a start. But still, we need to collect more information and expand our data set to find more intricate relations between urban entities in the scope of further urban pattern explorations which can help us reveal the inner nature of the traditional city shaped through history. 6. Conclusion The proposed study presented a method based on Data Mining technique for an exploration of urban patterns in the traditional city. Experiments carried out in this paper should be preliminary for further studies; therefore, a small part of Amasya- the traditional Hatuniye Neighborhoodwas chosen for the implementation. In the framework of the study, first, a database was prepared in GIS tools. Later, the dataset was used for Data Mining to investigate patterns and relationships among urban entities. Through this, transforming raw data into the data set and into useful knowledge about urban characteristics was aimed. The process of constructing the building database is still takes place, but obtained results in this paper showed that Data Mining presents various useful techniques to analyse raw urban information. For instance, numeric attributes can be classified according to its function, and

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

we can determine lower and upper size limits of urban entities. Also, the repetitive urban patterns can be detected and utilized by designers in the pre-design phases. Number of interpretations can be increased by collecting more data from the city and repeated Data Mining techniques according to the aim of the designers. The proposed study focuses on topographical and morphological attributes in the neighborhood. But in the future studies, demographic, economic and social data can be added to further analyse the traditional city patterns concerning socio-economic and cultural aspects of the city. Besides, because of the context-sensitive nature of the Data Mining results, the proposed method can be easily used in different databases created for different cities. As Bechhoeffer (2001) mentioned, urban culture and history are embodied and frozen, especially in the traditional cities. In order to sustain the continuity of the past, we must protect the physical, cultural and social character of the city. By this way, we can transfer the knowledge from the past and melt the past, present, and future of the city in the same pot to protect the urban culture and enhance local values which creates our lives in the first place. This study can help to constitute a rationale for the new design processes in traditional environments and also sustain the continuation of the urban patterns and characteristics in the city by revealing hidden knowledge of the city form. References Alexander, C. (1965). A City is not a Tree. Architectural Forum, 122(1-2), 58-62. Alexander, C., Neis, H., Anninou, A., King, I. (1987). A New Theory of Urban Design. New York; Oxford: Oxford University Press. Bechhoeffer, W., Yalçın, A.K.(1991). Amasya, Turkey: Lessons in Urbanity. Architect 40: Architecture in Development. September (1991), pp. 24-29, London: Concept Media Ltd. Bechhoeffer, W.(2001). Amasya: The Future of Tradition. In Turgut, H., Kellet, P. (Eds.), Traditional Environments in a New Millennium: Defining Principles and Professional Practice (pp.5154). Ankara: Yapı Endüstri Merkezi.

Carmona, M., Heath, T., Oc, T., Tiesdell, S. (2003). Public Places-Urban Spaces: The Dimensions of Urban Design. Amsterdam; Boston; London: Architectural Press. Duarte, J., P., Gonçalo, DS., Caldas, L.G. (2006). An Urban Grammar for The Medina of Marakech: Towards a Tool for Urban Design in Islamic Contexts. In Gero, J.,S. (Ed.) Design Computing and Cognition’06 (pp. 483-502). Springer: Netherlands. Ester, M., Kriegel, H. P., Sander, J., Xu, X. (1996). A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. In Simoudis, E., Han, J., Fayyad, U. (Eds.) Proceedings Book from The Second International Conference on Knowledge Discovery and Data Mining (pp.226231). AAAI. Fayyad, U., Uthurusamy, R. (2002). Evolving Data Mining into Solutions for Insights. Communications of the ACM, August (2002), vol.45, No.8, 2831. Fayyad, U., Piatetsky-Shapiro, G., Symth, P. (1996). From Data Mining to Knowledge Discovery in Databases. AI Magazine, Fall (1996), 37-54. Gil, J, Montenegro, N., Beirao, J.,N., Duarte, J., P. (2009). On the Discovery of Urban Typologies: Data Mining the Multidimensional Character of Neighborhoods. In Çağdaş, G., Çolakoğlu, B. (Eds.), Proceedings of the 27th Conference on Education of Computer Aided Architectural Design in Europe (pp. 269-278). ITU: Istanbul:Turkey. Gürbüz, E., Çağdaş, G., Alaçam, S. (2010). A Generative Design Model for Gaziantep’s Traditional Pattern. In Schmitt, G., Hovestadt, L., van Gool, L. (Eds.), Proceedings of the 28th Conference on Education of Computer Aided Architectural Design in Europe (pp. 841-849). ETH Zurich: vdf Hochschulverlag AG. Han, J., Kamber, M., Pei, J. (2011). Data Mining:Concepts and Techniques (3rd. ed.). Amsterdam; Boston; San Francisco; CA: Elsevier; Morgan Kaufmann. Hand, D., Mannila, H., Smyth, P. (2001). Principles of Data Mining. Cambridge, Massachusetts : The MIT Press. Hanna, S., (2007). Automated Rep-

Interpretation of urban data in the complex pattern of traditional city: The case of Amasya

resentation of Style by Feature Space Archetypes: Distinguishing Spatial Styles from Generatives Rules. International Journal of Architectural Computing, vol.5, issue 1, 1-23. Karabağ, K. (2010). A Computational Architecture Methodology For Design in Traditional Tissue: The Case Of Kalkan (Unpublished Ph.D. Thesis). METU; Institute of Science, Ankara. Kostof, S. (1991). The City Shaped: Urban Patterns and Meanings Through History. London: Thames and Hudson. Laskari, A., Hanna, S., Derix, C. (2008). Urban Identity Through Quantifiable Spatial Attributes: Coherence and Dispersion of Local Identity Through the Automated Comparative Analysis of Building Block Plans. In Gero J.,S., Goel, AK, (Eds.), Design Computing and Cognition’08: Proceedings of the Third International Conference on Design Computing and Cognition (pp. 615-634). New York: Springer. Miller, J., H., Han, J., (2009). Geographic Data Mining and Knowledge Discovery: An Overview. In Miller, J., H., Han, J. (Eds.), Geographic Data Mining and Knowledge Discovery (pp. 1-26). London, New York: CRC Press: Taylor and Francis Group. Özdemir, C. (1996). Amasya. Arkitekt. 9(96). No. 441, pp. 28-35, İstanbul: Nokta Basın A.Ş. Reffat, M.,R. (2008). Investigating Patterns of Contemporary Architecture Using Data Mining Techniques. In Muyyle, M. (Ed), Proceedings of the 26th

Conference on Education of Computer Aided Architectural Design in Europe (pp. 601-608). Antwerpen, Belgium: The Higher Institute of Architectural Sciences. Sohtorik, A. (2016). A Knowledge Discovery Approach to Urban Analysis: Beyoğlu Preservation Area As A Data Mine (Unpublished Ph.D. Thesis). ITU; Institute of Science, İstanbul. Sohtorik, A. S., Çağdaş, G., Sarıyıldız, S. (2010). Exploring the Patterns and Relationships of Urban Attributes by Data Mining. In Schmitt, G., Hovestadt, L., van Gool, L. (Eds.), Proceedings of the 28th Conference on Education of Computer Aided Architectural Design in Europe (pp. 873-881). ETH Zurich: vdf Hochschulverlag AG. Türkoğlu, E. (2006). Amasya İli, Hatuniye Mahallesi Geleneksel Yerleşim Dokusunun Analizi, Değerlendirilmesi Ve Koruma Geliştirme Önerisi (Unpublished Master Thesis).Gazi University; Institute of Science, Ankara. Yalçın, A.K.(1998). The Ottoman House Construction in the Old Town of Amasya. In Ireland, S., Bechhoefer, W. (Eds.), The Ottoman House: Papers from the Amasya Symposium (pp.9094). Ankara: The British Institute of Archaeology; The University of Warwick. Zaki, M.J., Meira, Jr. W. (2014). Data Mining and Analysis: Fundamental Concepts and Algorithms. Cambridge: Cambridge University Press.

ITU A|Z • Vol 15 No 1 • March 2018 • P. Çalışır Adem, G. Çağdaş

ITU A|Z • Vol 15 No 1 • March 2018 • 39-51

A model for parameterization of urban regulations

Miray BAŞ YILDIRIM mbas@itu.edu.tr • Department of Architecture, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.82574

Received: October 2017 • Final Acceptance: November 2017

Abstract The recent developments in digital design tools enable not only efficient but also holistic approaches to urban design problems. As the amount of data increases that urban design handle, the tools become more complex to maintain the spatial continuum across scales. In this context, the development of parametric design practices allows the generation of alternative scenarios and testing processes in addition to dealing with a large quantity of information. With reference to urban scale design problems, abovementioned information is closely related to regulatory processes, which mostly consist of text-based documents with 2D representations. This paper explores the use of parametric regulation modeling by presenting a design support model. The proposed model enables the designer to try out alternative scenarios by manipulating parameters and to gather real-time data, which is generated by using the Esri CityEngine software package. The textbased regulations are transformed into parametric form-based components by the proposed parametric regulation-modeling tool. In this regard, local regulations and standards are used for the generation of data-rich parametric 3D models and the evaluation of the alternative design scenarios. The first section provides a summary of the parametric regulation modeling and design guides within the context of urban design problems. The following parts describe the generation of the model as a parametric urban regulation model to support decision-making activities during design. Keywords Parametric regulation modeling, Urban design, CityEngine.

1. Introduction Urban design is usually located between architecture and planning scales, oscillating and filling a gap (Carmona et al., 2010). Schmitt (2012) describes design research approaches as simulations at different scales within the context of cities, which are building scale (small, S-scale), urban scale (medium, M-scale) and territorial scale (large, L-scale). These different scale problems are interconnected, as output from one can be input to another. Moughtin et al. (1999) distinguish the design cycle for regional planning from the design cycles of town planning, urban design and building design by identifying a feedback from the appraisal stage to the analysis stage, rather than from the decision stage. The relations between the tiers are not necessarily linear or unilateral for digital collaborative working environments. Furthermore, various design activities in built environments take place at different scales from a unit to region with the main actors of individuals to governments. The recent digital developments in design processes enable holistic approaches to urban design problems. As the amount of data increases, the tools become more complex to maintain spatial continuum with the transformation and analysis of data. The information modelling tools support such holistic approaches at different scales of design practices. Building information model (BIM), geographic information systems (GIS) or city information model (CIM) can be classified under such complex tools that enable data-rich interactive 3D environments, in addition to running various simulations. Urban scale design problems are related to several kinds of data, which can be gathered by GIS (visual) and regulatory processes (textbased documents). In a design process, designers analyze and synthesize such data, which emerges from variety of outputs in various formats. From the perspective of education, novice designers need to manage various kinds of data from different disciplines while building their design knowledge. In order to support novice designer’s design processes, an interactive data embedded model is proposed. The proposed

model allows designers to generate and test various design scenarios in a limited studio time. In doing so, regulations and standards are used to guide and gather real-time data. Novice designers encounter regulations and standards in terms of the representation of regulations (definitions) and evaluation of their design processes. The regulations and standards are mostly text-based descriptions with the complementary visual representations. Parametric approaches support the transformation of text based descriptions into form-based components to explore alternatives rapidly, as novice designers need the skills of greater critical thinking that is supported by experiential knowledge. Especially at urban scale design problems, the regulations are significant components to shape built environment. The politics of regulations or how regulations shape the built environment is beyond the scope of this study. This study addresses the regulations, as a tool to guide development and achieve desired design solutions, which includes the definitions and calculation methods to evaluate design scenarios by using both spatial and numerical outputs. Esri CityEngine software package is used to generate this model, which is a procedural modelling tool for urban space generation. As a city information modelling software, it provides a parametric design environment for users to generate and manipulate their scenarios. As an urban design problem, the scope of this study is urban regeneration areas, which constitute a major urban design issue in the cityscapes around the world. On the positive side, urban regeneration enable designers to work with an entire and existing neighborhood as a whole. The model is designed to be adaptable for various urban regeneration design cases. The paper describes the structure of the proposed model as a parametric regulation-modeling tool to support design processes. The first section below provides a summary of the parametric regulation modeling and design guides within the context of urban design problems. The following parts describe the generation of the model, which

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

involves the transformation of regulations into parametric forms as an educational decision support tool. 2. Parametric urban regulation modeling Regulations are important components of the urban regeneration processes. Design codes and regulations can be considered as systematic precedents in addition to being guides to fulfill requirements. Design strategies, regulating plans, master plans are occurred to “to guide the development process and achieve a range of enhanced design outcomes, while all provide a clear two- or three-dimensional vision of future development form” (Carmona et al., 2010, p.313). Lawson (2004) points out legislators as one of the design knowledge constraints. He put an emphasis of the requirement of the great amount of knowledge to set criteria with the satisfaction of standards and attach these attributes to designs. Lawson (2004) also criticizes the regulations as they are generated for easy measurements rather than what is desirable. However, from the perspective of education, novice designers need to understand regulations and standards and the related concepts and definitions. Design codes or design guidelines can be considered to be more flexible than the regulations. Some are motivated by user participation. Earlier examples are mostly based on defining urban morphologies. Also they can be described as systematic approaches to urban design problems, which emerge from the precedent cases (Kolodner, 1993). One of the most well known early definitions of code is Christopher Alexander’s Pattern Language, which is first published in 1977 as a systematic organization of urban morphology (Derix, 2012). “Responsive Environments: A manual for designers” is another practical book with an aim to introduce design ideas “that the built environment should provide its users with an essentially democratic setting, enriching their opportunities by maximizing the degree of choice available to them” (Bentley et al., 1985, p.9). More recent example is the Smartcode, which is described as “a form-based code that A model for parameterization of urban regulations

incorporates Smart Growth and New Urbanism principles” (Transect, n.d.). Smartcode is not a digital platform but it is a model to guide design principles. Design guidelines/design codes are not used in definition of the proposed model, however they are considered as potentials for future studies. More flexible environments like parametric urban design aids to understand and generate sustainable and livable urban environments with the knowledge of urban form and its data. Parametric regulations modeling aids the design process to observe the manipulation of parameters with the outcomes of numeric values like floor area ratio, built are, number of units etc. Modelur, CityCAD, CityZoom, LandXplorar are commercial parametric regulation modeling tools. Some of them can collaborate with other programs like GIS and some of them are standalone programs (Figure 1). Derix (2012) points out they all share the common characteristic of working from a designed masterplan, which includes determined circulations, movement structures and blocks. In other words, the plot development regarding the massing is the common characteristic of such tools. In addition to that as the focus is on parameters for regulations, the detail of models can be limited. Modelur allows for parameters such as building height, gross floor area and floor numbers to be set and for numerical values such as building area and floor area ratio to be observed (Modelur, n.d.). At the same time, it offers various types of land use with real-time calculation reports (Modelur, n.d.). Although it is known to be an easy tool to use software with online accessibility, its visualization potentials are limited. CityZoom is a system that “integrates several performance tools that allow the simulation of different attributes related to a planned or existing city” (Grazziotin et al., 2004, p.216). It serves as a platform to operate various building performance models with the aim of optimizing urban planning process (Grazziotin et al, 2004). CityZoom can work with different kinds of data such as freehand draw-

Figure 1. (a) CityCAD (CityCAD, n.d.), (b) Modelur (Modelur, n.d.), (c) LandXplorar (LandXplorar, n.d.), (d) Cityzoom (Grazziotin, 2004, p.218).

ings or aerial images in various file formats. It is used for the simulation of potential buildings with urban regulations (Grazziotin et al, 2004). Users can define building regulation related values (setback values, add floor number, floor height, etc.) by an urban regulation editor and also they can observe the numerical outcome of their alternatives as well as their model’s visual representations. CityZoom can be defined as a powerful tool to visualize regulation limitations and observe the related numeric data with the defined parameters. However, building detail levels and the lack of other components (like streets, open spaces) only aids to evaluate generated alternatives as the masses of buildings. CityCAD, “urban design software tool for conceptual 3D master planning of sites” (CityCAD, n.d.), is similar to Cityzoom and Modelur in being a commercial product for the early stages of urban design and allows sustainability, livability and viability analysis. LandXplorer Studio Professional is a product of Autodesk. It is described as an interactive, real-time authoring system that allows, “to visualize 3D geospatial data and to effectively create, analyze, manage and distribute geospatial information” (LandXplorar, n.d.). It integrates with GIS data, 2D Autocad data, 3D models, legends and textures (LandXplorar,

n.d.). It can be used in various scales such as large-scale city models or neighborhood scales. Even though, it presents a flexible environment, customization of rules and detailed mass structures cannot be defined. The parametric regulations modeling tools that are described above are used for 3D master planning purposes with real-time data gathering. However, they have limited customization properties to make calculations to analyze. Regarding that the proposed model differs from the parametric regulation modeling tools by the guidance of local regulations and standards and integration of different typologies to generate 3D models with detailed visualizations. 3. CityEngine as a city information modeling tool Esri CityEngine is a well-known procedural modeling software package that enables parametric design. By using CityEngine, designers can quickly create virtual environments that fulfill visualization of required functions. On the other hand, by manipulating rule-defined parameters, generated models can be manually transformed. The procedural logic in modeling cities depends on L-systems and generation of two-dimensional patterns, which is achieved through the application of shape grammars (Parish & Müller,

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

Table 1. The main components of the model.

2001). The integration with ArcGIS geo-database files is one of the main capabilities of the software, which enables users to use required urban data. Through real-time visualization, it gives opportunity to observe reactions of certain moves on manipulating parameter values both as visualization and data gathering. Recent examples on CityEngine enable users to observe performance of buildings or regulatory processes. Two of these examples are Redlands Redevelopment plan (Esri Redland Redevelopment plan, n.d.) and Auckland City Plan (Esri Auckland City Plan, n.d.), which supports users to observe different scenarios. Especially Auckland City Plan project is used to generate alternative scenarios for different regulatory acts of governments (Esri Auckland City Plan, n.d.).

Figure 2. The model’s workflow. A model for parameterization of urban regulations

4. The transformation of regulations and standards into the proposed model The model is aimed to be a responsive tool for novice designers with data rich 3D visualization to alter between different design ideas by manipulating parameters. The transformation of regulations and standards into rules are one of the main issues in creation of the model. Accordingly, ‘Standardized Building and Zoning Regulations’ (SBRZ) (T.C. Çevre ve Şehircilik Bakanlığı, 1985), ‘Regulation on Preparation of Spatial Plans’ (RPSP) (Resmi Gazete, 2014), ‘Dimensioning and Design Principles of Urban Roads’ (DDPUR) (for street generation) (Turkish Standards Institution, 1992), “Cars – Design Criteria of Auto Parking Facilities in Urban Areas” (Turkish Standards Institution, 2013) and “Auto parking regulations” are used in the model as a guide to generate alternative scenarios. As regulations and standards consist of text-based descriptions, they cannot be fully transformed into rules due to program limitations, semantic conflicts or some regulation exceptions. The model has three main components based on SBZR and the operating logic of CityEngine, which are built environment (facilities and buildings), networks (streets) and open spaces (Table 1). Basically the model consists of the computer generated architecture

(CGA) rule files, Python scripting modules and the components like images 3D objects to generate the 3D visualization (Figure 2). Rule files are assigned to shapes to create models in which default values are assigned as rule-defined values. From the inspector menu, predefined parameter values can be observed and be manipulated by sliders. After generating a 3D model of an area, existing model’s reports can be extracted by using the python module like population, density or household number. 4.1. Buildings The building rule consists of two types of buildings, which are residential/commercial buildings and facility buildings. The common parameters for all building types consist of the model display, topography adjustment and facade construction. The model display includes mass, envelope and detailed display of the model. Different colours are assigned to mass display mode, based on the code of usage type defined in RPSP. Users can decide the function of a building by selecting the building usage type from parameters. The components of a 3D model of a building is described under three parts regarding the related regulations and standards, which are parcel limitations, building envelope and building parameter. 4.1.1. Parcel limitations The model’s all rule files have the properties of parcel limitations based on SBRZ. Accordingly, four main zones are defined in SBRZ, which are ‘commercial and residential areas’, ‘industrial areas’, ‘small industrial areas’ and ‘except residential urban areas’. As mentioned beforehand, the focus is limited to residential urban regeneration the zone type of ‘commercial and residential areas’ is constituted in detail. The basic parcel limitations consist of the parameters of the parcel width and the parcel depth. Based on SBRZ parcel width and depth minimum values are given in Table 2 and 3. While applying a building rule to a shape, the rule checks whether it fulfills minimum required values, which consists of the values of parcel depth and width based on the zone type. If the values

Table 2. Parcel limitations – Parcel depth.

Table 3. Parcel limitations – Parcel width.

are not fulfilled the requirements, the green areas are generated by the rule. Other than urban zone type, the existence of front yard is also a parameter in parcel depth definition (Table 2). Parcel width limitations depend on building’s settlement that includes being corner parcel or being attached or detached building type. Also, building floor numbers are another determining parameter defined in rules (Table 3). According to related regulations, front yards are defined whether they exist or not. Therefore, the rule has two cases, which are ‘without frontyard’ and ‘with frontyard’, users can choose either one of them.

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

Table 4. Set back values extracted from SBRZ.

Figure 3. The envelope with L-shape footprint.

4.1.2. Building envelope The limitations of building construction area are described in the envelope definition, which includes the information of setback distances based on building floor numbers. According to SBRZ, the envelope of a building consists of the extrusion of setback limitations. The minimum setback values related to building floor numbers and maximum heights are displayed in the Table 4. In the model, the setback values are appeared as rule-defined values unless the designer needs to change manually. The setback values, building floor numbers, ground floor height and upper floor heights are the parameters that affect the envelope, in addition to being interconnected. Another paA model for parameterization of urban regulations

rameter for the building construction is being detached or attached building that are defined based on regulations. In rule, depending on the choice of building type “attached” or “detached”, the side setbacks set as being active or passive. The envelope option can be observed by selecting model display mode as ‘envelope’, which presents building with semi-transparent envelope (Figure 3). 4.1.3 Building parameters Basically building parameters consists of two types of buildings, which are commercial/residential buildings and facilities. When a building rule file assigned to a shape, building floor numbers are given randomly between the values of 1-4 floors, up to 4 floor is defined to be a default value. Users can also choose between the options of 4-9 floors and 10 and more, which changes the parcel width limitations as well. Residential and commercial buildings are defined in the same rule file. Population data is extracted from the residential building’s definition by calculation of the values of gross floor area and average floor space per person. Technical and social infrastructures are defined as facilities in general, which are classified as educational, health, accommodation, social and cultural, religious and public facilities based on regulations. Color codes for density and building functions and walking distances that enable visual evaluations in addition to understanding and using legal color codes. In addition to defined facilities, multi-storey parking garage is defined under facilities CGA rule file. Based on TS-10551, multi-storey parking limitations set as parameters such as maximum-minimum capacity of cars, floor numbers, ground floor heights and upper floor heights in addition to calculation of car capacity. The walking distance is generated as a transparent circle with a radius data, which is assigned as walking distance value. By walking distance, users can visually observe walking distances to certain facilities/green areas, which consists of a transparent circle with radius parameter as a distance value (Figure 4). Based on the type of the

facility walking distances defined as: playground 500 m, high school 1500 m, open sports area 500 m, kindergarten 500 m, elementary school 500 m and secondary school 1000 m. Users can observe the minimum legal requirements of walking distances by the predefined values. Also, the walking distance values can be manipulated by selecting custom mode. The legal color representations support novice designers to observe the legal colors of zoning and density. Different colors are assigned to mass display mode depending on their usage type based on RPSP legend of usage of areas also same colors are used in the envelope display mode (Table 5). Another legal color representation is the density, which is calculated from the total data. 4.2. Networks (streets) The streets are generated based on the “Dimensioning and Design Principles of Urban Roads TS-7249” (DDPUR). Street construction rule consists of two parts. The first part is the general part, which includes texture, sidewalk, green area, bicycle, crossing and roadside parking parameters (Figure 5). Users can determine the existence, values and location of these components. For instance, parameters of crosswalks consist of width, location (start or end of the street) and distance from start or end of street. Bicycle road parameters consist of the existence of bicycle road (on/off) and being whether the one way or two way bicycle road. The dimensions and the color of the bicycle road are gathered from the regulations of the designing and building the bicycle ways, bicycle stations and bicycle parking in urban roads. Based on that, minimum total width of a oneway bicycle road is 160 cm including 50 cm safe distance and dividing strips. The minimum two-way bicycle road width is defined 270 cm as default value. In addition to that the color blue is assigned to the bicycle road based on related regulations. Crosswalk as an important component to determine human traffic defined in the street construction. The parameters of crosswalk consist of the width of crosswalk, the location of crosswalk (start or end of

Figure 4. Walking distance circle and its parameters. Table 5. Building functions and their designated colors based on RPSP.

a street) and the distance from start or end of a street. The width of crosswalk is defined as 2.5 meters depending on the definition of the minimum crosswalk width for residential area urban roads in the DDPUR. On the other hand, users can change the value by manipulating the crosswalk width parameter. The second part contains the parameters of street generation depending on type. The streets are generated based on the “Dimensioning and Design Principles of Urban Roads” (DDPUR). DDPUR presents the minimum lane widths related to urban road types, which consists of four main urban road types: orbital road, urban district connection road, urban district collector road and urban district road. Based on

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

Figure 5. The street parameters.

the road types, lane widths are changed also can be observed in reports section. Each road type has the parameters of street type (divided/undivided), lane parameters (lane number, lane width divided/undivided), refuge (texture green/concrete, width) and service road (left and right, on/off). Roadside parking limitations are also defined in this part based on â&#x20AC;&#x2DC;Cars â&#x20AC;&#x201C; Design Criteria of Auto Parking Facilities in Urban Areasâ&#x20AC;&#x2122;. Depending on that standard, 5 main parking types are defined (Figure 6). Also, since the dimensions of different parking are defined, generated roadside parking car capacity is calculated within the CGA rule file. Calculated car capacities are reported in the same name with other parking capacity data for exported data reports.

Figure 6. Roadside parking types. A model for parameterization of urban regulations

4.3. Open areas The open areas consist of green area and parking lots, which can be selected after applying the open space CGA file to a shape. Open parking lots have the same definitions with the green areas in terms of the division of the parcel but with different texture. The data that is gathered from the generated open parking area consists of the total area

Figure 7. Open sports area with the parameters.

of the parking space and the car capacity of the generated area. In order to calculate car capacity 28 m²/car is used as an approximate value, which is gathered from related regulations. The numeric data of green spaces and their number can be observed in the reports section once the whole area is selected. Also, under the definition of open areas, football, tennis and basketball areas inserted to the project with the standard sizes, which supports novice designers to generate 3D model with the correct dimensions rapidly (Figure 7). 4.4. Analysis of generated scenarios: Calculations based on regulations and standards The real-time reports that stem from the generated models can be observed under two ways. The first one is the report data from the selected 3d models, which are defined under CGA rule files named as primary data. The values of the floor area ratio, gross floor area, ground area ratio, parcel width, parcel depth and parcel area for building construction are gathered by primary data generation. In addition to that for residential areas, the data of population, unit numbers and generated car parking numbers are also defined in these rule files. The population data helps to generate secondary data such as gross residential density, unit number and the demand of car parking place. The calculation of

population depends on the type of the residential building as single-family housing or multi-family housing. For single-family housing and pre-defined typologies Parker Morris Report’s (1961) minimum dwelling size values are used to calculate population (Towers, 2005). For the others, the average floor per person value is used to calculate population, which is set as 30 sqm per person. Since this value is interchangeable depending on the location, it is defined as a parameter to be manipulated. The unit number can be calculated by population and Turkey’s average household value which is 3,6 people/unit (2014’s data) from Turkish Statistical Institute. This value used as an input to calculate the demand of car parking space in the secondary data gathering process. The calculation of car capacities and the needed car space is related to each type of rule file since different type of data is extracted to calculate these values. The definition of rules of parking lots is referenced from Turkish Standards Institute’s “Cars – Design Criteria of Auto Parking Facilities in Urban Areas (TS10551)” and “Auto parking regulations”. Depending on standards, two main types of auto parking are defined. The first one is building parking lots (closed or open) the other one is public parking facilities. In this study, especially public parking facilities are defined which consists of two types, which are roadside parking and off-

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

Table 6. The requirements of facility areas depending on population. Simplified from RPSP.

Table 7. The second type of reports.

street parking. The reports section includes the value of number of needed car space and the number of generated car space. The value of the generated car space is calculated in three different components. The first one is the multi-storey parking garage in the facilities rule file. The second one is open parking lot defined in open space construction. The last one is roadside parking defined in street construction. Underground parking garage is not defined. For comparison of the need of cars and existing car capacities, the A model for parameterization of urban regulations

required number of auto-parking is determined by the rules taken from regulations as: For residential every 3 units, commercial areas for 50 m², accommodation facilities, the number of parking space changes depending on type like hotels every 5 room; pensions every 4 room; hostels every 5 room. For social/ cultural/sports facilities every 30 seating, education facilities 400 m², health facilities 125 m², religious facilities 300 m², government facilities 100 m². For accommodation facilities and social/ cultural/sports facilities room number and seating number can be given manually. The others can be calculated from area value or unit value that is described above. The second way of gaining report data is to export report data with the python script as text file, which helps to make calculations for the total area. The data from rule-generated reports are used as input to maintain these calculations. In order to that, two python scripts should be used after the generation of whole area. The first script works as getting total values from the generated area as a text file. The second python script calls back this text file and based on the gathered values, makes calculations. The outcome of these calculations can be defined under two parts, the first part consist of the existing situation like the values of the total parcel area, total population, household number, density, generated parking space, green area per capita and generated facilities’ values. Density value can be assigned to shapes as an object attributes which aids to represent density by righteous colors defined by regulations in density display mode. Depending on RPSP, technical and social infrastructure minimum area needs are defined according to population density. The population intervals are defined as 0 -75000, 75001-150000, 150001500000, 501000 and more. As the study area is limited to neighborhood scale, the part from a city, the default values are given according to minimum population as guidance for novice designers. The related values are presented in Table 6. Based on population data that is gathered from the model (calculated from residential buildings) the ap-

proximate minimum requirements for technical and social infrastructures are calculated by running python script. These report data is used to generate minimum necessity of facilities. The second part consists of the requirements calculated from the values from the former calculations, which are the demand of car parking space, demand of the facilities and their minimum requirements. Table 7 presents the second type of data gathering for overall evaluation. These calculations aid novice designers to observe the numeric outcomes of their design decisions and to reevaluate them if it is needed in addition to stand by their design while it is evaluated. 5. Conclusions The complex design problems of urban regeneration, characterized by administrative processes as much as physical and socio-economical transformation, require multidisciplinary approaches and participatory processes. The proposed model, a responsive design tool built on the procedural logic of CityEngine, introduces some means for a collaborative, not yet participatory, process to be of use in the educational context. Users can generate and test different design scenarios by receiving visual and numeric outputs on the spatial qualities. The real-time data supports decision-making processes by allowing designers to understand and become aware of the implications of their design scenarios. The definitions of the parameters and calculations are emerged from the local regulations and standards, which are used to guide novice designers in order to be informed rather than to limit them. Since novice designers need to manage large amount of data in a limited time frame, the proposed data integrated model supports their design process to have a deeper understanding of the relevant concepts, while building their design knowledge. The model differs from other parametric regulation tools by customization of local regulations to evaluate design scenarios. Moreover, as an educational platform, it enables novice designers to familiarize with the terms and legal representations (color codes,

walking distances). The transformation of regulations includes not only the definition of parameters, but also the analysis components such as the walking distance feature. The information that emerges from the standards supports the design process by rapid generation of components with the correct dimensions like dimensions of different kinds of car parking space or standard sizes of open sports area. However, the proposed model has several limitations regarding the transformation of all regulations and standards into the model due to technical limitations. Another limitation is the lack of the components of urban circulation system, which is planned to be resolved with further studies. The structure of model enables the integration of improvements that consist of different analysis scripts or pre-defined components. For future studies, the model can be transformed into a participatory tool with the support of the future design codes and guidelines. References Bentley, I., Alcock, A., Murrain, P., McGlynn, S. & Smith, G. (1985) Responsive Environments: A Manual for Designers. Butterworth Architecture. Carmona, M., Heath, T. Oc, T. & Tiesdell, S. (2010) Public Places, Urban Spaces: The Dimensions of Urban Design: Architectural Press/Elsevier. CityCAD (n.d.). Retrieved from https://www.holisticcity.co.uk/index. php/citycad. Derix, C. (2012) Digital Masterplanning: Computing Urban Design. In Urban Design and Planning: Institution of Civil Engineers, Thomas Telford Publishers, pp. 203-217. Esri Redlands Redevelopment Plan. (n.d.). Retrieved from http://www.esri. com/~/media/Files/Pdfs/library/brochures/pdfs/cityengine-example-redlands.pdf. Esri Auckland Unitary Plan. (n.d.). Retrieved from http://www. algim.org.nz/globalassets/ symposium-gis/2013-gis-symposium/speaker-presentations/auckland-council-3denablement-of-auckland-unitary-plan. pdf. Grazziotin, C. G., Turkienicz,

ITU A|Z • Vol 15 No 1 • March 2018 • M. Baş Yıldırım

B., Sclovsky, L., & Freitas, C.M.D.S. (2004). CityZoom- A Tool for the Visualization of the Inpact of Urban Regulations”, Sigradi, 216-220 Kolodner, J. L. (1993). Introduction. Case-based reasoning. Morgan Kaufmann, New York. ISBN 1-55860237-2. LandXplorer. (n.d.). Retrieved from http://download. autodesk.com/us/ landxplorer/docs/ LDX11_Studio/index.html Lawson, B. (2004) What Designers Know, Elsevier-Architectural Press, Oxford. Modelur. (n.d.). Retrieved from http://www.modelur. com/home. Moughtin, C., Cuesta, R. & Sarris, C.A. (1999) Urban design: Method and techniques. Oxford, England: Architectural Press. Parish, Y. I. H., & Müller, P. (2001) Procedural modeling of cities. Paper presented at the SIGGRAPH ‘01 Proceedings of the 28th annual conference on Computer graphics and interactive techniques, Los Angles. Resmi Gazete (2014). Mekansal Planlar Yapım Yönetmeliği. Retrieved

A model for parameterization of urban regulations

from http://www.resmigazete.gov.tr/ eskiler/2014/06/20140614-2.htm Schmitt, G. (2012) A Planning Environment for the Design of Future Cities. Stefan Müller Arisona, Gideon Aschwanden, Jan Halatsch, PeterWonka (Eds.) Digital Urban Modeling and Simulation Communications in Computer and Information Science Volume 242, 2012, 3-17. SideFX, (n.d.). Houdini. Retrieved from http:// www.sidefx.com. T.C. Çevre ve Şehircilik Bakanlığı. (1985). Planlı Alanlar Tip İmar Yönetmeliği. Retrieved from http://www.csb. gov.tr/turkce/index.php?Sayfa=sayfa&Tur=mevzuat&Id=144 Transect (n.d.). SmartCode. Retrieved from http:// transect.org/ codes.html. Turkish Standards Institution. (1992). Şehiriçi Yollar- Otolar için Otopark tasarım kuralları (TS-10551). Turkish Standards Institution. (2013), Şehiriçi Yollar Boyutlandırma ve Tasarım Esasları (TS-7249). Towers, G. (2005) An introduction to urban housing design: At home in the city. Oxford: Architectural Press.

ITU A|Z • Vol 15 No 1 • March 2018 • 53-64

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions Ömer Halil ÇAVUŞOĞLU1, Gülen ÇAĞDAŞ2 1 omerhalilcavusoglu@gmail.com • Architectural Design Computing Graduate Program, Graduate School of Science, Engineering and Technology, Istanbul Technical University, Istanbul, Turkey 2 cagdas@itu.edu.tr • Department of Architecture, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.80488

Received: October 2017 • Final Acceptance: November 2017

Abstract Over the last decades, technological advancements were carried out with a great pace and that forced industries to drastic changes and paradigm shifts. These advancements provide new opportunities that arise with their new requirements. Due to some of these requirements, AEC industry unwittingly caused some crucial global issues which are gaining momentum exponentially, cannot be ignored anymore. The main reason of this situation is identified as many significant decisions which directly affect the performance of the building and the relationship of the building with natural and built environment are taken, even if there is no certain and valid information. The focus of the study is to discuss and evaluate the collected data and the obtained findings from previously implemented 5 case studies with 25 unique participants in a same context to re-evaluate and understand how BIM can help designers in the early stages of architectural design, most particularly in decision making processes. In addition, we also focus on investigating what opportunities it provides, what drawbacks it causes and what the user feedbacks about using the tool in these stages are. The focus of this study is not to offer an alternative way for traditional design practices but to explore if these kinds of tools have advantages for conceptual designing and/or design supporting. To achieve these aims, we have used quantitative (questionaire), qualitative (pure observation, participant observation, in-depth interviews and focus groups) and protocol analysis (retrospective analysis) methods. Keywords BIM, Decision making in design, Early design stage, Performance evaluation.

1. Introduction With the beginning of the Industrial Revolution and its subsequent World Wars, mechanical and electrical systems were invented and developed with a great pace and achieved technological advancements forced industries to drastic changes and paradigm shifts. While this situation provided new opportunities, it also directly affected the requirements of the era. Due to some of these requirements, an expeditious increase in AEC industry has started which has caused some crucial global issues in progress of time such as global warming, depletion of ozone layer, depletion of natural resources and so on. Researches show that AEC industry is directly responsible for this situation. For instance, all consumed energy of the United States used up by the buildings is 47,6% (Architecture 2030, 2017), while it is 40% in EU countries (The European Union, 2012). In the United States, when all other manmade immovable structures are included - things such as bridges, roads, dams, and ports - the raw materials consumed in the process of construction exceeds 75 percent of the total (Roodman et. al, 1995; Matos and Wagner, 1998). From another point of view, the construction in the United States consumes three times more raw material than all other economic and industrial activities combined (Smith and Tardiff, 2009). Furthermore, the population of the world is expected to reach to almost ten billion by 2050 (United Nations, 2017). In a similar vein, during next twenty years, it is expected to be more than doubling the built environment (Krygiel and Nies, 2008). It is obvious that demanding for materials and energy is growing exponentially. On the other hand, according to the United Nations, the world can barely sustain a population of six billion at a middle – income consumption level, which we experienced in 1990s (United Nations, 2005; Smith and Tardiff, 2009). These facts show us that the world will not be able to meet these exponentially growing infinite demands in a short span of time. This situation leads us to be in a much more critical position which cannot be ignored anymore.

Numerous researches have shown that one of the most important reasons of this consumption is many essential decisions (building orientation, building shape, structural system, building envelope and so on) are taken at the early design stages without any valid and certain information (Gervasio et al., 2014; Granadeiro et al., 2013; Hong et al., 2000; Holm, 1993; Gratia and De Herde, 2003). These decisions which are taken with often inadequate information on the site, climate, geography also provide a basis for the final performance and the aesthetics of the final outcome. In order to find a solution to recent challenges, new BIM processes and capabilities have started to be developed and although they have been acknowledged as inadequate for the early stages of architectural design, they can have a huge potential to support these stages with their existing and potential capabilities. The focus of this study is to discuss and evaluate the collected data and the obtained findings from the previously implemented 5 case studies with 25 unique participants in the same context to re-evaluate and understand how BIM can help designers in the early stages of architectural design, most particularly in decision making processes. In addition, we also focus on investigating what opportunities it provides and what drawbacks it causes, and what the user feedbacks about using the tool in these stages are. The focus of this study is not to offer an alternative way for traditional design practices but to explore if these kinds of tools have advantages for conceptual designing and/or design supporting. To achieve these aims, we have used quantitative (questionaire), qualitative (pure observation, participant observation, in-depth interviews and focus groups) and protocol analysis (retrospective analysis) methods that provided us to obtain a wide range of data. This study consists of five main sections. We firstly review the the decision making processes in early design stages and its significance over the entire design process and the final product. We also review the literature on the early stages of architectural design and its significance within the scope of de-

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

sign cognition field. Later on, we concentrate on the BIM and discuss what capabilities BIM can offer for the early stages of design and how these capabilities can help us to challenge global issues that we are going to face. In the Case Study Implementations section, we present the previously implemented case studies. In that section we also express the research methods that have been used during the case studies and explain how we use these inputs and outcomes for this specific study. In the Findings section, we explain and evaluate the collected data of the previous implementations and discuss the findings of different studies in the same context to constitute an understanding for the focus of the study. In the last section, we discuss the findings of this study connection with the related literature in terms of decision making, building information modeling, energy efficient design and so on. 2. Decision making in early design and the potential benefits of BIM 2.1. Decision making in design Decision-making can be described as a four-step process: searching the environment for problems, analysis and development of possible courses of action, choosing a particular course of action and implementation of the action (Sprague 1980; Simon 1960). It is a process choosing a preferred option or a course of actions from between a set of alternatives based on given criteria or strategies (Wang et al., 2004; Wilson and Keil, 2001) by identifying, gathering information, and assessing alternative resolutions. From the design perspective, as design problems are usually open-ended, complicated, uncertain and ambigious, decision making process in design has a dynamic nature. Dynamic decision making is characterized by the following four features: a series of decisions is required to reach the goal, the decisions are interdependent, the state of the decision problem changing a consequence of the decision maker’s actions and the decisions to be made in a real-time environment (Edwards, 1962; Brehmer, 1992). Dynamic decision making environment also requires two overlapping cognitive activities which track key

variables for information, regarding present and expected conditions and the control, generation, evaluation and the selection of alternative actions (Lerch and Harter, 2001). Within this context, feedback is one of the most crucial features of dynamic decision processes. There are three decision support mechanisms that are commonly accepted and discussed in this regard (Gonzalez 2005; Arora, 2009): • Outcome feedback refers to providing decision makers with feedback on the performance results of their decisions. • Cognitive feedback refers to giving decision makers instructions on how to perform the decision task. • Feedforward refers to providing decision makers with an environment to perform what-if analysis of potential decisions. 2.2. Early design stages Design is identified as a process of generating new, valuable and desirable solutions by many researchers (Casakin, 2008; Woo, 2005; Buchanan, 2001; Galle and Kroes, 2014; Galle, 2011). It is also acknowledged as one of the most complicated cognitive process of human beings (Liu and Architecture Group, 1996; Akin, 1979; Oxman, 1996; Gero and Mc Neill, 1998) which is mostly accepted as a problem solving activity. Design problems usually have open-ended, wide-ranging, complicated and ambiguous characteristics (Pinch et al., 2010; Carmel-Gilfilen and Portillo, 2010). They are interacted with various criteria to generate design ideas and concepts (Casakin, 2008; Goldschmidt, 1989). Goel (1995) states that design problems are constituted of lots of different parts and elements, and these do not need to be logically connected. This makes design a cyclical process which is fed from the feedbacks of the ongoing process. To deal with these complicated and ill-defined problems, Alexander (1964) claims to separate design problems into smaller sub problems. Thus, it will be possible to obtain more defined problems which can be solved successfully with rationality (Goldschmidt, 2014). In these stages, designers are also expected to decide on significant factors

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

such as building orientation, building shape, structural system, building envelope and interior finishes with inadequate and indefinite information. These decisions which are taken with often inadequate information on the site, climate, geography, also provide a basis for the final performance and the aesthetics of the final outcome. This situation makes early design stages important not only for the aesthetic and functional aspects but also for sustainability. In order to find a solution to this problematic situation and enhance early design processes, some researchers and practitioners have focused on developing new design tools and information management databases for meeting the new requirements of the current practices. In this context, BIM comes to the forefront with its current and potential capabilities. 2.3. BIM and the potential benefits for early design stages National Institute of Building Sciences (n.d.) defined BIM as a model and process: “A building information model is a digital representation of physical and functional characteristics of a facility. As such, it serves as a shared knowledge resource for information about a facility forming a reliable basis for decisions during its life cycle from inception onward. A basic premise of Building Information Modeling is collaboration by different stakeholders at different phases of the life-cycle of a facility to insert, extract, update or modify information in the model to support and reflect the roles of that stakeholder”. Building Information Model is a data-rich, object-based, intelligent and parametric digital representation of the facility, from which views appropriate to various users’ needs can be extracted and analyzed to generate feedback and improvement of the facility design (GSA, n.d.). This integrated, parametric, intelligent and object-based virtual model provides an environment to obtain abstract forms of representations, inferences, work and time schedules, analyzes, simulations, and so on. By this way, BIM enables the creation, generation and management of all numer-

ic and non-numeric data coordinately, collaboratively, coherently, synchronously, and in a computable way from the conceptual design to the end of the building’s life (Krygiel and Nies, 2008; Garber, 2009; Deustch, 2011; Eastman et al., 2011). The characteristics and capabilities of BIM are discussed with many reports and publications in both industry and academic settings (Eastman et al., 2008; Azhar, 2011; buildingSmart, n.d.; AGC, n.d.). In spite of the fact that BIM offered unique capabilities during the drafting and construction processes, it was regarded as being inefficacious as an early design environment for a short time before. Particularly for the past 10 years, BIM has been able to meet the capabilities of the other CAD tools commonly used in the industry and has reached a level of competitiveness with them for early stages of architectural design. Today, BIM stands out with its ‘intelligent and parametric modeling capabilities’ and ‘simulation, analysis and inference capabilities’ when comparing with conventional CAD tools in terms of early stages of architectural design process. As it was mentioned in the previous sections, the intent of developing BIM as an early design environment is directly related to the recent global challenges which AEC industry are obligated to face. Studies show that the difficulty of evaluating the performance of the building in the early stages of design has made the products generated to be inadequate in achieving the desired performance (Schlueter and Thesseling, 2009). Regarding studies which investigate how to achieve more sustainable design approaches, the necessity of evaluating sustainability criteria from the initial stages of design has been determined (Attia and De Herde, 2011). With BIM, even in the early stages of architectural design, designers can begin to work on models to analyze and redesign their designs according to the performance and efficiency criteria using specific pre-defined presumptive data without requiring any advanced engineering knowledge. Since many important building-related decisions have not yet been taken at the early design process, the capabil-

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

ity to analyze and simulate the geometric (mass) design models of the BIM tools is a distinguishing feature, which is one step ahead of conventional design tools. In this context, the passive and active strategies identified by Krygiel and Nies (2008) for the sustainable BIM approach are reduced and sorted according to the possibilities BIM can provide in the early stages of architectural design: building orientation, building massing, solar and shadow analysis, daylighting analysis, conceptual energy modeling, potential renewable energy analysis. In brief, it is emphasized that taking advantage of essential information about design task in early design process is useful and crucial. BIM environment with its information processing capability operates as an improved design support tool with powerful drafting and modeling features, performance simulations and visual analysis feedbacks. These feedbacks are beneficial as visual and numeric outcomes of the design which enable the evaluation of the designed mass to improve it. Then, the design relies on the functional, aesthetic, sustainable realities and the subjective judgements of designers. In this section, we have studied the current realities of early stages of architectural design and building information modeling by literature review, and then, we have explained why and how we need BIM in these stages. 3. Case study implementations Over the last years, we have implemented 5 case studies with 25 unique participants to investigate the opportunities and drawbacks of BIM environment in the early stages of architectural design. Participants were consisted of undergraduate and graduate students of various architecture education programs of Istanbul Technical University. Some of these case studies have been published (Cavusoglu, 2015a; Cavusoglu, 2015b; Cavusoglu and Cagdas, 2017). All participants indicated that they were familiar with at least one wellknown CAD tool to use it as drafting, modeling and designing tool. On the other hand, they had no or little experience about using BIM and energy

modeling. For this reason, since participants had no or little experience about BIM related concepts which would be used in these studies, a series of theoretic and practical lessons had been given to the participants. We started these lesson series with giving a lecture about general concepts of BIM and its importance for AEC industry. At that moment, our focus was not about using BIM in early stages of architectural design but describing why BIM gain importance in a short span of time in AEC indusrty and what opportunities it provides. After informing the participants about these general concepts, we started to give applied courses for using BIM particulary for basic operations such as drawing and modeling geometric forms and also understanding the non-geometric features of the environment. At the last sessions of these lesson series, we concentrated on using BIM in early stages of architectural design. While we were focusing on mass modeling and conceptual energy analysis features of the environment, we also gave theoretical courses about the main principles of sustainable building design.By this way, we tried to provide an adequate basis for participants to achieve a successful research process within a limited time. Design tasks of these implementations had different complexities which varies from ‘designing a single function building form’ to ‘designing a multi-purpose functional building which has functional, contextual and sustainable requirements that must be considered’. To achieve the expected results, we have used quantitative (questionaire), qualitative (pure observation, participant observation, indepth interviews and focus groups) and protocol analysis (retrospective analysis) methods that led us to be able to obtain a wide range of data. During these studies, we have examined the BIM in terms of efficiency evaluation of different roles of the tool, software evaluation criteria and design cognition, most particulary how using BIM in early design stages affects decision making processes of the designers. On the other hand, in consideration of the previous case study implementations, we took a step back and re-eval-

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

uate all these processes in this study. We reconsidered all the collected data and the obtained findings from these studies to study and discuss all together within the same context. 4. Findings 4.1. Efficiency evaluation of BIM’s roles in early design stages During the case study implementations, we determined 5 main roles for the BIM environment that contribute to the early design stages. These are design exploration, 3D modeling, parametric modeling, energy modeling and decision support system. The participants were asked to vote these capabilities from 1-10 points through their own experience and to explain the reason why they had evaluated the capabilities in this way. The average results for each role are given below in Figure 1. Design Exploration. The participants have a consensus on that BIM environment is not as good as traditional sketching and physical modeling for design exploration and triggering creativity. On the other hand, they also indicated that starting the design process within sketching environment and then improving the concept design in BIM is a very effective way of working without being obligated to sacrifice any important aspects of early design stages. It is understood that BIM environment offers lots of unique capabilities that supports the designer in early design stages in comparison with other CAD tools but still cannot compete with sketching environment in terms of design exploration. The main problem we observe in the case is that the analyzing capabilities of BIM led the participants to improve their designs they developed before. From this moment on, we notice that the design process is starting to evolve from an intuitive desing approach to a systematic design approach. In other words, after starting the analyzing and evaluating process in BIM, almost none of the the participants tried to develop an entirely different concept models. So it can be discussed that using BIM in creative design process may cause fixiation so that the participants tended to focus solely on one solution and tried to improve it.

Figure 1. Graphical representation of questionnaires average results.

3D Modeling. Almost all users found the BIM’s capabilities equal or better than the traditional CAD tools for 3D modeling. But, while we were investigating the participants’ feedbacks on 3D modeling capabilities of the tool, we found a contradictory situation that some participants voted very high, but others voted low. When we thoroughly investigated this situation, we understood that the participants who had a better command with the tool, were able to do what they wanted easily which directly led them vote higher. On the other hand, the participants who were not able to use the tool fluently voted lesser. Parametric Modeling. Other than the geometric modeling, the participants stated that they took advantage of working with a model which constitute its elements parametrically connected. In regard to our observations, this situation actualises in two ways: BIM provides an environment where users are able to constitute a parametric connection on 2D/3D models while they can still manipulate the model manually. In this way, it offers a way of working which provides an environment that users can work totally manual, parametric, and also parametric with manual transformations. In one of our implementations, participants were able to continue their design development processes with these parametric design capabilities of the tool. Mostly,

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

they tackled with the problem of constituting true parametric connections while most of them had no parametric and algorithmic thinking foundation. Because of this reason, we ignored the parametric design capabilities of the tool for this series of implementations. On the other hand, in all of our implementations, participants used the parametric foundation of the environment for taking inferences via calculations and schedule capabilities such as floor area schedules classified by different functions, building masses and zones. All participants found this useful and expressed that it gave them a chance to handle the geometric and non-geometric inputs much more easier and effective. Conceptual Energy Modeling. Almost all participants even the ones who were experienced some troubles using the BIM found conceptual energy modeling as the most powerful and useful capability of the environment. We observe that conceptual energy modeling directly contributes two important factors. Firstly, it supports the designers not only with its text based and visual analysis feedbacks to obtain better design solutions in terms of performance output but also provides them a self learning environment to design sustainable solutions with the way of trial and error problem solving method. Decision Support System. Firstly, we met the same situation here in common with the evaluation of 3D modeling that some of the participants voted higher than the others voted lesser. The reason is again same with the 3D modeling evaluation that is directly related with the command of the tool. On the other hand, based on our observations and the participantsâ&#x20AC;&#x2122; feedbacks within the interviews and focus group meetings, BIM environment provides a lot of capabilities to its users in terms of decision supporting. But more importantly it has a suitability to be developed as a better decision support system. Over the last years, there have been a lot of new tools which have been dedicatedly developed for enchancing decision support processes in early design stages. This has made BIM environment much more efficient within this perspective. Along with this, we

notice that the participants voted fairly higher in the subsequent implementations when compared with the first ones. Overall Evaluation. As a summary of the criteria we found out that BIM could not be able to provide a good environment for design exploration. Moreover, it needs more flexibility for 3D modeling capabilites in terms of ease of use and user friendliness which were started to be developed nowadays. Having a parametric foundation serves great with both geometric and non geometric inputs of the process. In addition to being parametric, conceptual energy modeling capability provides a great environment to test and evaluate the developed design model without any expert engineering knowledge. All these capabilities provide a solid foundation for BIM in terms of being a decision support system. Recently, new capabilities and tools have been developed to take BIMâ&#x20AC;&#x2122;s decision support role a step further which aims to enhance cognitive processes of designers while they are analyzing and evaluating the design models. It is important to underline that even the BIM tools have new opportunities for the early design stages, we still have to educate and improve ourselves to be able to adapt to this new approach. As it is a well known general issue for BIM, we also experience and observe that the need of adopt the cultural shift for BIM is a must to get efficiency from it. In early stages of design, this cultural shift is not only about improving the command of the tool but also gaining knowledge about related concepts, improving computational thinking and analytic reasoning skills. During the implementations, even it was not our target to observe BIM as a learning environment, we determine that the participants who lack command with the sustainable design principles were using the tool with trial and error method to understand which analysis input and output affected the design in what way. By this way, they figured out the lack of command issue and constitute a new way of working for themselves. This situation shows us that BIM provides an effective environment for learning, teaching and

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

practicing the sustainable design. In a similar vein, BIM not only enhances the analytic reasoning and evaluation skills of designers, but also provides an environment for improving them. 4.2. Efficiency evaluation of BIM environment in terms of software evaluation criteria In this section, we discuss how the participants evaluated the environment in terms of software evaluation criteria. We determined 8 criteria for evaluating the environment which are capabilities, developability, ease of use, effectiveness, flexibility, functionality, learnability and online education sources. The participants were asked to vote these criteria from 1-10 points through their own experience and to explain why they evaluated the criteria in this way. The results as an average for each role are given below in Figure 2. Almost all the participants indicated that the varied capabilities of the environment are useful and helpful. This situation directly contributes a possitive effect to effectiveness and functionality of the environment. The participants stated that they were able to analyze and evaluate some design parameters which they were obligated to ignore in their early design activities. Moreover, they expressed that there are also a lot of potential capabilities which can be developed to make BIM a much more efficient environment for early design stages. The main point of their remarks were focusing on the capabilities based on new analyzing and evaluation features which can enhance the environment particulary in for decision making aspect. On the other hand, some of the criteria relatively has lower points in our questionaire. In interviews and focus group discussions, it is understood that some of the participants were not able to gain enough command with the tool. They underlined that tool had to show progress in terms of the ease of use and flexibility. In addition, they asserted that the online education sources were not enough in both quantitative and qualitative manner which directly affected the learnability of the tool. However, from the first day of our implementations up to today, follow-

Figure 2. Graphical representation of questionnaires average results.

ing the popularity of using BIM based capabilities in the early design stages, there has been an immense progress to overcome these mentioned and also unmentioned issues. Educational institutes and professional organizations were initiated online education programs to develop designers’ technical, theoretical and practical knowledge of BIM and sustainability in early design stages. Furthermore, new BIM based environments and capabilities are being developed for analyzing the concept model with different aspects and helping the designers’ evalution processes with new features. Within this context, we detect that while the powerful features of the environment are still being developed, the weaknesses are also being developed to a much more better position. In a similar vein, we observe that the participants who participated the subsequent implementations marked higher votes when compared with the previous ones. Based on our observations, interviews, focus group discussions and the results of the questionaire, we detect that the environment provides very powerful capabilities for the early design stages. In addition, the basis of BIM provides an environment where a lot of other tools or features can be developed. For that reason, we think BIM will continue to gain importance for early design stages especially as a decision support system which enhances designers’ cognitive activities with supporting them its capabilities.

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

4.3. Evaluation of decision making processes in BIM environment Kerstholt and Raaijmakers (1997) discusses that the decision maker needs to have an accurate model of the related elements and their temporal characteristics. By this way, they can control and predict the current and future states of the process which enable both feedback and feedforward mechanisims. With the capability of object oriented modeling where elements are parametrically connected to each other with involving all numeric and non numeric information, BIM provides an environment to obtain feedbacks and feedforwards from the design model through computing all the related information and serving them as a meaningful outcomes such as text based and visual based analysis reports and simulations. By this way, BIM provides a basis to the participants where they track the information about their key variables of design task interpretations and use it to handle, generate, evaluate and select between the design alternatives. It is also accepted that decision making has 6 important steps to follow which are constructing the problem, compiling the requirements, collecting the information, comparing the alternatives, considering the factors and commiting to a decision and improving it. We also observe that BIM environment is able to support designers in all these phases with its aforementioned capabilities. From another perspective, Kleinmuntz (1985) distinguished two different strategies which are action-oriented strategies and judgement-oriented strategies. Action-oriented strategies are used as decision makers who apply their actions and observe their effect on the system and proceed depending on the observed effect. On the other hand, judgement-oriented strategies are used as decision makers first try to reduce the uncertainty of the problem by requesting information and then apply their actions. During the implementations, we experienced that the participants were mostly using action-oriented strategies. They were taking decisions, applying actions and observing their effect on the model and proceeded improving design by this way. In contrast to this situation, based

on our observations, we think that BIM is also suitable for the judgement-oriented strategies, especially if the design problem could be structured and framed as more rational subproblems. In conclusion, we want to underline that 5 different case study implementations with 25 unique participants and a specific BIM environment are not sufficient to generalize the outcomes. Moreover, as most of the users had no or little experience with BIM and sustainable design principles but just learned the basics of them for that specific implementations, it is not fair to evaluate their votes as a precise input. Associated with this, as we have discussed before, we notice that the participants who have better command with the tool, tended to vote pretty higher than the others. It is obvious that we need to carry out further implementations with more participants who have better proficieny with the tool and sustainable design principles to comprehend better the current situation of the tool in early design stages. But still, the findings show how well BIM environment provides a foundation for early design stages, also where it is facing problems and what the underlying reason of this problem is. 5. Discussion and conclusion The main objective of the study is to observe and explore what opportunities BIM can offer in early design stages particulary in decision making processes. During the study, we also investigate how the participants evaluated BIM as an early design stages environment in terms of the roles of BIM and software evaluation criteria. Due to the ongoing devastating incidents, considering the performance criteria in the early stages of design to achieve a better performative buildings become obligatory. The perspective of Foqué (2010) accords with this situation as he expresses that ‘intuitive thinking and rational thinking are not oppenents; they are the twin poles between which the artist structures reality’. He also asserts that architecture must take advantage of both science and art (Foqué, 2011). As technology has evolved over the years, lots of new digital design tools

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

have been started to be developed with a great pace. With the emergence of new technologies, new design approaches and way of working opportunities are shaping the way we think, make decisions and design. From this perspective, BIM does not only signify an environment or a tool but also a way of working. BIM’s ability to store, inference and analyze the data related to the building serves as a modeling and decision support environment for designers in early design stages. In addition to its decision support capabilities, it offers a real time object oriented modeling environment where all the design model parametrically connected to each other with all numeric and non numeric inputs. By this way, designers can continue to design while they are considering both aesthetic, functional and also sustainable factors within a cyclical design process (Azhar et al., 2009). Within this context, BIM comes to the forefront with its current and potential capabilities as a decision support system for early design stages. Kymmell (2007) describes the basic concepts of human action and interaction as being interwoven with each other as visualization, understanding, communication and collaboration, and explains how the direct and indirect features of BIM have fed these four concepts. It is now possible to consider many different factors in the early design stages which are often overlooked or difficult to assess nowadays. Kymmell explains the contribution of BIM through the phrase “a picture is worth a thousand words” as follows: “then how much will a 3D model be worth, or a movie of a timed sequence of events?” In a similar vein, Miller (1956) has shown that short-term memory of individuals is limited to 7 ± 2 elements during data processing. Because of the assumption that design moves made during a problem-solving action are the representations of the minds at that moment they are considered to be connected to each other in the range of 7 ± 2 moves (Goldschmidt, 2014). Thus, as Kymmell points out, the building information model and all the inferences derived from it help designers to search for a more effective design by playing a reminder and decision sup-

port role for many design criteria that are prone to ignoring in the conceptual design process and disappearing from the working memory. In conclusion, BIM offers many benefits to its users not only for drafting and construction processes but also for early design stages. It may have a way to become a better design tool and a decision support system. But, as we have discussed before, it already has very powerful capabilities that enhances decision making processes particularly for sustainable expectations of design. References AGC. (n.d.) The Contractors’ Guide to BIM. http://www.agc.org. Akin, O. (1979). Exploration of the design process. Design Methods and Theories, 13 (3/4), 115-119. Alexander, C. (1964). Notes on the Synthesis of Form (Vol. 5). Harvard University Press. Arora, H. (2009). Building decision support for dynamic decision making: A design science approach. Arizona State University. Attia, S., De Herde, A. (2011). Early design simulation tools for net zero energy buildings: a comparison of ten tools, In: Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association, Sydney, Australia, pp. 94–101 Azhar, S. (2011). Building information modeling (BIM): Trends, benefits, risks, and challenges for the AEC industry. Leadership and management in engineering, 11(3), 241-252. Azhar, S., Brown, J., Farooqui, R. (2009). BIM-based sustainability analysis: An evaluation of building performance analysis software. In Proceedings of the 45th ASC annual conference (Vol. 1, No. 4, pp. 90-93). Architecture 2030. (n.d.) http://goo. gl/o4FC5X. Brehmer, B. (1992). Dynamic decision making: Human control of complex systems. Acta psychologica, 81(3), 211-241. Buchanan, R. (2001). Design research and the new learning. Design issues, 17(4), 3-23. buildingSmart. (n.d.) www.buildingsmart.com.

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

Carmel‐Gilfilen, C., Portillo, M. (2010). Creating mature thinkers in interior design: Pathways of intellectual development. Journal of Interior design, 35(3), 1-20. Casakin, H. (2008). Factors Of Design Problem-Solving And Their Contribution To Creativity. Open house international, 33(1). Çavuşoğlu Ö. H., Çağdaş, G. (2017). Why Do We Need Building Information Modeling (BIM) in Conceptual Design Phase?. Çağdaş, G., Özkar, M., Gül, L. F., & Gürer, E. (Eds.). Computer-Aided Architectural Design: Future Trajectories: 17th International Conference, CAAD Futures 2017, Istanbul, Turkey, July 10-14, 2017, 121-135. Çavuşoğlu, Ö. H. (2015a). Building Information Modeling Tools: Opportunities for Early Stages of Architectural Design. CAADRIA 2015 - 20th International Conference on Computer-Aided Architectural Design Research in Asia: Emerging Experiences in the Past, Present and Future of Digital Architecture, 427-436. Çavuşoğlu, Ö. H. (2015b). The Position of BIM Tools in Conceptual Design Phase: Parametric Design and Energy Modeling Capabilities. eCAADe 2015 Volume 1 Real Time, 607-612. Deutsch, R. (2011). BIM and integrated design: strategies for architectural practice. John Wiley & Sons. Eastman, C. M., Eastman, C., Teicholz, P., & Sacks, R. (2011). BIM handbook: A guide to building information modeling for owners, managers, designers, engineers and contractors. John Wiley & Sons. Eastman, C., Teicholz, P., Sacks, R., & Liston, K. (2008). BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers and Contractors. Edwards, W. (1962). Dynamic decision theory and probabilistic information processings. Human factors, 4(2), 59-74. Foqué, R. (2010). Building knowledge in architecture. ASP/VUBPRESS/ UPA. Foqué, R. (2011). Building knowledge by design. Texto apresentado na IV Jornadas Internacionales Sobre Investigación En Arquitectura Y Urbanismo.

Galle, P., & Kroes, P. (2014). Science and design: Identical twins?. Design Studies, 35(3), 201-231. Galle, P. (2011). Foundational and instrumental design theory. Design Issues, 27(4), 81-94. Gero, J. S., & Mc Neill, T. (1998). An approach to the analysis of design protocols. Design studies, 19(1), 21-61. Garber, R. (2009). Optimisation stories: The impact of building information modelling on contemporary design practice. Architectural Design, 79(2), 6-13. Gervásio, H., Santos, P., Martins, R., & da Silva, L. S. (2014). A macro-component approach for the assessment of building sustainability in early stages of design. Building and Environment, 73, 256-270. Goel, V. (1995). Sketches of thought. MIT Press. Goldschmidt, G. (1989). Problem representation versus domain of solution in architectural design teaching. Journal of Architectural and Planning Research, 204-215. Goldschmidt, G. (2014). Linkography: unfolding the design process. Mit Press. Gonzalez, C. (2005). Decision support for real-time, dynamic decision-making tasks. Organizational Behavior and Human Decision Processes, 96(2), 142-154. Granadeiro, V., Correia, J. R., Leal, V. M., & Duarte, J. P. (2013). Envelope-related energy demand: A design indicator of energy performance for residential buildings in early design stages. Energy and Buildings, 61, 215-223. Gratia, E., & De Herde, A. (2003). Design of low energy office buildings. Energy and Buildings, 35(5), 473-491. GSA, GSA BIM Guide. (n.d.) https:// goo.gl/3Vamrd. Holm, D. (1993). Building thermal analyses: what the industry needs: the architect’s perspective. Building and Environment, 28(4), 405-407. Hong, T., Chou, S. K., & Bong, T. Y. (2000). Building simulation: an overview of developments and information sources. Building and environment, 35(4), 347-361. Kerstholt, J. H., & Raaijmakers, J. G. (1997). Decision making in dynamic task environments. Decision making:

Enhancing decision making processes in early design stages: Opportunities of BIM to achieve energy efficient design solutions

Cognitive models and explanations, 205-217. Kleinmuntz, D. N. (1985). Cognitive heuristics and feedback in a dynamic decision environment. Management Science, 31(6), 680-702. Krygiel, E., Nies, B. (2008). Green BIM: Successful Sustainable Design with Building Information Modeling. Wiley, Hoboken. Kymmell, W. (2007). Building Information Modeling: Planning and Managing Construction Projects with 4D CAD and Simulations (McGraw-Hill Construction Series): Planning and Managing Construction Projects with 4D CAD and Simulations. McGraw Hill Professional. Lerch, F. J., & Harter, D. E. (2001). Cognitive support for real-time dynamic decision making. Information systems research, 12(1), 63-82. Liu, Y. T., & Architecture Group. (1996). Is designing one search or two? A model of design thinking involving symbolism and connectionism. Design Studies, 17(4), 435-449. Matos, G., & Wagner, L. (1998). Consumption of materials in the United States, 1900–1995. Annual Review of Energy and the Environment, 23(1), 107-122. Miller, G. A. (1956). The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological review, 63(2), 81. NBIS, Building Information Modeling (BIM). (n.d.) http://www.wbdg. org/building-informationmodeling-bim. Oxman, R. (1996). Cognition and design. Design studies, 17(4), 337-340. Pinch, S., Sunley, P., & Macmillen, J. (2010). Cognitive mapping of creative practice: A case study of three English design agencies. Geoforum, 41(3), 377387.

Roodman, D. M., Lenssen, N. K., & Peterson, J. A. (1995). A building revolution: how ecology and health concerns are transforming construction (pp. 11-11). Washington, DC: Worldwatch Institute. Schlueter, A., & Thesseling, F. (2009). Building information model based energy/exergy performance assessment in early design stages. Automation in construction, 18(2), 153-163. Simon, H. A. (1960). The new science of management decision. Smith, D. K., & Tardif, M. (2009). Building information modeling: a strategic implementation guide for architects, engineers, constructors, and real estate asset managers. John Wiley & Sons. Sprague Jr, R. H. (1980). A framework for the development of decision support systems. MIS quarterly, 1-26. The European Union: Directive 2012/27/EU of the European Parliament and of the Council of 25 October 2012. (2012). Official Journal of the European Union no. 55. doi:10.3000/19770677.L_2012.315.eng United Nations. (2017). World Population Prospects 2017. https://goo. gl/48ujS6. United Nations. (2005). Population Division of the Department of Economicand Social Affairs, World Population Prospects: The 2004 Revision. Wang, Y., Patel, S., Patel, D., Wang, Y. (2003). A layered reference model of the brain. In Cognitive Informatics, 2003. Proceedings. The Second IEEE International Conference on (pp. 7-17). IEEE. Wilson, R. A., & Keil, F. C. (Eds.). (2001). The MIT encyclopedia of the cognitive sciences. MIT press. Woo, H. R. (2005). Creative Abilities in Design. The International Journal of Creativity & Problem Solving, 15(1), 101-113.

ITU A|Z • Vol 15 No 1 • March 2018 • Ö. H. Çavuşoğlu, G. Çağdaş

ITU A|Z • Vol 15 No 1 • March 2018 • 65-77

An investigation on algorithm aided BIM approaches to increase collaboration and optimisation in project phase: A case study Şeymanur YILDIRIM seymanur.yildirim1007@gmail.com • Department of Informatics, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.44711

Received: October 2017 • Final Acceptance: November 2017

Abstract There is a need for new workflow that comes with Algorithm Aided BIM to be used and modified by different users within business association. If efficiency of new workflow is valid only limited users, automation provided in the project doesn’t fully ensure the reduction of stress because of extensive labor and limited time. For this purpose, it is aimed to create a cooperative working environment for users, who have different programing knowledge level, in the new workflow like BIM. In the present case, there aren’t comprehensive studies because of that it is aimed to prepare an approach plan for standardisation of new workflow in order to spread the usage among the employees and to provide comprehensive optimisation in the project phases. The approach is tested with a case study to show the effects on a project phase. As a result of the analysis, it is concluded that algorithms reduce the tension arising from the heavy workload in the projecting process but for the comprehensive optimisation, algorithm is not sufficient. Algorithm should be standardized to increase the company-wide usage, reduce the dependence of users on each other and pauses, caused by lack of information. In the absence of standards, in spite of being able to optimize, it is monopolized by the limited user and it is difficult for new users to understand and adapt. Therefore, in addition to the algorithms, used to reduce stress due to heavy workload and time loss during the projecting phase, the use of a common language or standards are necessity for companies. Keywords Algorithm aided BIM, BIM, Parametric modelling.

1. Introduction Building Information Modeling (BIM) has become widespread with supplying integrated environments for different disciplines. One of the pivotal goals of BIM environments is to build a virtual construction model of the building by incorporating everything into a single source model. Depending on object-based logic, BIM consists of multidimensional digital representations and features of various facilities. The objects used in BIM programs are smart components that contain data and parametric rules. By means of the database used in BIM programs, relations with different objects can be established. In this view, building components can be modified in a single place, allowing differentiation in all dimensions. Since all phases of a project can be executed through a single model, BIM provides opportunities to organize workflows that will increase efficiency in the project phase. Moreover, problems that can be encountered in the design and construction process can be predicted from the beginning and resources can be managed correctly and controlled by the stakeholders. One of the common problems in the context of Computer Aided Design (CAD) and Computer Aided Manufacturing (CAM) approaches to architecture offices today is the lack of communication between different disciplines. As a result, conflicts can arise between different actors and problems could be more dramatic because of causing a costly change in the progressive stages of the project. However, advancing by adopting the integrated design approaches from the first stage of the project process, problems that may be encountered in the future can be reduced. As a result, BIM supports both the design and construction phases of the project and offers better analysis (for the building life cycle), control and testing as opposed to produce by conventional CAD environments. The increasing level of complexity and the data load in computer aided design models such as vast number of parameters, variables, assumptions, relations among geometries necessitate a new perspectives to handle the complexity.

Aish (2013) approaches CAD in three eras: the 2D drafting era of early 1980s, BIM era consisting of 2D drawings and 3D models of buildings and later the design computation era (Aish, 2013:5; Humppi and Österlund, 2016). Aish’s term of design computation era points out multidimensional generative and relational 3D building models which are coded and executed by using data through graphs or scripts. Generating a 3D building model through graph or scripts provides opportunities for merging design, analytical and fabrication models in one model and also design automation. In other words, different phases of a project can be represented in one model through parameters and relations. Investigations on ways for expanding the boundaries of BIM towards design computing led to led to the emergence of the concept of Algorithm-Aided BIM (Humppi and Österlund, 2016). Further to the theoretical discussions on the integration of algorithmic approaches and BIM, today in practice as well, it is possible to generate 3D building models in BIM environment based on constraints, rule sets, and requirements of project to achieve more efficient solutions. It can be considered as the reflection of object-oriented low-level programming logic onto computer-aided design representation. In a broader sense, depending on the efficiency of the algorithms, the efficiency and overall performance of the project phases can be improved. However, in this case it becomes important to investigate what kind of algorithms and at which extend the algorithms improve the efficiency in the project phases. For example, when the designers do not have experience in coding, would traditional way of using BIM become more efficient? In the scope of this research, collaboration of different disciplines through Algorithm Aided BIM approaches is investigated. Moreover, the potentials and limitations of Algorithm Aided BIM are discussed in the context of complex optimisations during the project phase. We argue that Algorithm Aided BIM approaches accelerate the adaptation of users to different levels of knowledge through algorithmic workITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

flow in collaborative design processes in which different disciplines involve. The preliminary results show that without standardisation in collaborative environment, optimisation merely coming with algorithmic approaches does not fulfil the overall efficiency expectations. Due to the fact that, standardisation of Algorithm Aided BIM project process is investigated for comprehensive optimisation. In the current situation, there is any standardization for the Algorithm Aided BIM project stages. The aim of this study is that standards for the new approaches are prepared with the help of the subway design project which is undergoing the design process and literature review. 2. Algorithm aided BIM Algorithm allows users to find a solution efficiently with his/her own recipe. In order to make the project process algorithmic, firstly, a problem needs to be identified, and the solution is abstracted according to the problem context. This abstraction is based on translating a nonlinear process into a linear logic, defining infinite number of possibilities through finite number of steps by using variables, functions, operations and relations. The relations involve organisation and automatisation of repetitive actions, iterations and recursions. Further to the constitution of algorithms, in most of the cases the variables are differentiated, and the solution is tested. This very basic description of algorithms can be applicable and elaborated in different contexts and cases. When it comes to BIM, certain associations allow the entire system to be managed by a limited number of users under network control with algorithmic approaches. Algorithms have potential to reduce the workload and time spent in the project development, particularly in repetitive tasks. Moreover,, since the entire system is related to each other, the data network can be visualised more clearly through visual scripting environment. Therefore, changes in the data network and the new design decisions become trackable by different actors BIM programs are based on the interrelationships between the objects controlled by the variables. Modifying

the variables give alternative model results. As expressed by Eastman et al. (2011), there is algorithmic infrastructure at BIM programs. While the entire system is under control in the algorithmic modeling, the object-based singular control is observed in BIM programs. Therefore, as the model variables become complicated, difficulties of adapting to changes and the number of errors is increasing. Because of that, the relationship between BIM and the Algorithmic Modeling is investigated and attracted much attention (Janssen, 2014). In algorithmic modeling, users develop a general logical scenario and controls the entire system. This gives the user a very comprehensive and automated process. However, in BIM programs, parameter control can not go beyond the object base, and systems in this context differ from algorithmic modelling. According to Boeykens (2012), many users use algorithmic modeling software to produce complex models and transfer them to BIM programs after model reaches a certain level of maturity. After the model transfers to the BIM environment, there is no return. Furthermore, for programs aiming to create a rich database like BIM, only the introduction of geometric representations from the early phase of design is causing serious information loss. Within the scope of this research, it is suggested that the user should adapt to algorithm aided approaches in order to reduce the workload, to control the process more and to design the possible events in a short time. Algorithm aided approaches provide users with the ability to automate tasks, generate parametric geometries, and giving ability using model to perform various engineering analyzes. Users are constantly working on different alternatives to reach the best solution. However, in order to achieve more efficient solutions, algorithm provides the programming of decisions based on constraints, rule sets and project requirements. This provides a transition from more traditional solutions to a wider range of possibilities. In traditional BIM project steps, users, who need to work intensively and

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

Figure 1. Factors that push the user in the process of BIM.

spend a lot of time, are searching new project steps. This study aims to examine which project steps might need more time and effort in the modeling process. These steps, at which algorithmic definitions become more efficient in comparison to traditional way of BIM usage, are defined as breakpoints. At the breakpoints it is considered that transformation of the BIM processes into algorithmic processes would be more advantageous. In the scope of the study, seven breakpoints are determined based on the individual experiences of the author and literature review. Controlling the data load and finding the required information from the desired layer are some of the crucial problems facing the user. These lead to breakpoint in the project workflow. As shown in figure 1 below, there are seven factors which cause the breakpoint. The increase in LOD (Level of Development), nD (dimension of model), the maturity level of the model, the number of disciplines, the complexity of the project, the scale differences, and the number of repetitions of similar operations push the user into new workflows because of increasing the number of data loads and variables in the model. More integrated systems are needed to strengthen the parametric infrastructure of BIM so that all expected operations can be done correctly and accurately at a limited time. 3. Application method for algorithm aided BIM BIM allows many disciplines to cooperate in a single database. In this way, all disciplines can follow and control the process from the first stage of

the project. As a result of the integration of algorithm aided approaches and BIM, new project steps need to provide possibilities to work within unity. Otherwise, there may be information loss between project phases due to the disunities. This study aims to prepare a cooperative working plan for different disciplines in the visual programming environment (algorithm aided environment). However, it is difficult to make a general plan that defines algorithm aided project steps because the method to be followed is various even though the goal is only one. While defining new workflow, it is aimed to present a proposal to give the users a general idea about algorithm aided project steps. The process can be converted algorithmic to solve the problem. However, it is also important to develop an algorithm in a standardisation to work in cooperatively because the established logic can be changed with the arrival of a new business decision. At this stage, changing the algorithm against the new problem may cause purpose to change. This could cause moving away from the solution, or even uncontrolled progression. It may also need efforts to reestablish the logic of the algorithm. As a result, other users need to be able to understand and interpret the flow of the algorithm when it is examined. For this purpose, it should be made orderly and understandable. A remarkable amount of study has been done to improve the usability and intelligibility of prepared algorithms. For this purpose, visual programming is developed to meet this need. Visual programming is an easier language than script language (Crafai, 2015). In visual ITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

programming, codes are also modular. When the user indirectly creates a program graphic, it generates programming codes hidden in the nodes of the graph. The level of complexity could be reduced with the algorithm being modularized. Moreover, according to a survey of 25 participants by Woodbury et al. (2007), designers have observed that the small changes in the algorithm have increased the clarity of the script. Woodbury et al. (2007) mentioned that it is necessary to group the script in modules that do similar works. In this way, scripts can be represented in groups so that the design intent is more clearly conveyed to the other users. In addition, for the grouping of patterns according to Alexander’s (1977) definition, it is necessary to clearly define the problem, the abstract solution and the results. Therefore, the process shouldn’t be limited to only grouping of modules. Woodbury (2010) also mentioned coloring scenarios. This coloring ideas was produced by him based on Christopher Alexander’s design patterns. It is colored by similar commands or problem solutions, making it easier for users to understand the pattern. Today’s object-oriented programming foundation reflects onto the way of creating logic in the graphical user interfaces as well. The objects incorporate their own design philosophy into their own pattern of their own worldview at the structural level within the group they belong to. The coloring scenarios have potential to improve clarity while coloring creates bridges between script, operations and the 3D building model. The standards related to visual programming which are examined in this study can be adapted to Algorithm Aided BIM tools. In the scope of the study one specific program was chosen to be used due to avoid potential differences caused by program limits. Standards are established according to the most widely used BIM tools in program selection. According to a study by BIPS (2014), the most commonly used software in architecture, engineering and construction was found Autodesk Revit program. Therefore, Algorithm Aided BIM is being implemented through Autodesk Revit which is preferred by

applications and standards generally. As a visual scripting environment, Dynamo which is an open source add-on for Autodesk Revit was used. On the other hand, it is possible to use Dynamo outside of or in Revit. Dynamo provides diverse possibilities for object creation, object modification and data management. In terms of expanding Revit’s capacities towards computational design tool, Dynamo enables customization in design processes. Currently there is no commonly accepted standard for the use of Algorithm Aided BIM. The work of White Research Lab has been observed, indicating that these tools are designed to promote wider use at the office regardless of the level of experience among employees (Ondejcik, 2016). According to the firm standards, the functional part of Dynamo is divided into two sections: the user interface and the background. The partitioning provides for the distribution among users with various visual programming skills. Users with very basic skills can interact with the interface without ever having to worry about ruining the backface function. The interface is supported by color coding to describe the functions of the algorithm. The bright colors are open to user interaction and the gray colored groups are related to the back face. The back face is used among the users, who have advanced level of knowledge about algorithm. It is considered that Algorithm Aided BIM environments need to allow different users with different programming knowledge level to be used. For this purpose, Christopher Alexander’s patterns are set out to show how different users would include an algorithm for classification and what kind of information is needed for a better understanding. Since there is no specific standard in the present case, limited amount of published standards were examined. In order to achieve a standardization in Algorithm Aided BIM, the following items are assumed to be defined as minimum requirements: • Group definitions and project description, • The purpose of the algorithm, • Algorithmic way of design descriptions,

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

Figure 2. Dynamo usage template.

• Problem should be prepared by grouping the algorithms for the solution purpose. The above mentioned 4 items are the logic of establishing the algorithm that should be standardized by the users at different levels apart from the the program focus. The standards to be created in the program are examined with the sample works produced in the project process. The advantages and deficiencies are tested and an alternative standard is proposed for project firms. A template was created as shown in Figure 2, using Woodbury’s (2010) algorithm to classify algorithms, Alexander’s (1977) idea of design patterns, grouping methods used in Grasshopper visual program software, and standards established by White Research Lab using Dynamo software. Groups are colored according to their functions so that different users can understand the algorithm and their functions are settled and their functions can be easily understood. In the information section (Figure 2), a brief summary of the algorithm is described containing the name of the person who prepared the algorithm and the information about which version of the software is used. In this way, other users would know who to consult when they want to get information about the algorithm. Specifying the information for each module of the algorithm reduces the information loss. Therefore, when a part of the code will be changed, the change history would be trackable. Specifying the version is crucial. Otherwise, the same code may not work due to changes or removals of functions in a short time.The information section is expressed in blue color. In order to increase the intelligibility, the blocks in the algorithm are grouped by their functions. This way of grouping also provides a better understanding for the users who do not have

programming experience. The gray color refers to the part that makes up the design, is defined as the background and is not open to all users. By differentiating the access permissions to the background code, the program code is protected from potential errors made by novice users. The green group is the modifier of the algorithm, which allows assigning new information on the objects. Invalid information seen in the database can be checked by the control groups because it is the last place the data was stored before it was written. Purple group includes geometry and other limitations. The orange group refers to the groups associated with the Revit program. It includes commands such as obtaining information from the project, selection procedures, writing information. The light blue group is the part used for data surveying in the algorithm. Commands such as data extraction from outside are included in this group. The dark blue groups refer to information parts. Algorithm descriptions, user information and writing information are included in this section. The pink groups refer to the code being studied. The necessity of the created template, its deficiencies and developments are examined by case study. 4. Assessment of standards with case study In order to reduce the tension caused by intensive workload and time-consuming during the project phase in BIM environment, it is aimed to establish an approach for the companies that want to switch their projects and workflow into Algorithm Aided BIM process. Creation object-based database in BIM makes it possible to work more efficiently and synchronously at every step of the project process in comparison with the conventional usage of BIM. Unlike the conventional ITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

methods, the users can focus on design decisions, rather than spending time on documentation. However, the integration of BIM with algorithm aided approaches can be achieved by the end of project planning, automation and optimization by consuming less labor and time. If multiple disciplines are considered to work at the same model in BIM, relatedly a similar approach needs to be followed in the case of the Algorithm Aided BIM. In order to test the validity of the assumptions on algorithm aided approaches in BIM, a case study was established. Users with different backgrounds in one construction company have attended to the case study. The following questions were investigated through the case study based on design and development of one subway project: • What can be the need for algorithmic approaches to BIM environment? • Is it possible for the project stakeholders involved in the project planning process to work in a coordinated and collaborative way in Algorithm Aided BIM processes? • How can users who have different programming knowledge levels adapt to the Algorithm Aided BIM process? • What method should be used in the process of model development for users who go to Algorithm Aided BIM processes? The reason for choosing subway project is that the author has a desire to benefit from professional practice and experience and is more dependent on the collaborative working environment and information exchange, which is important because of the important role of the other stakeholders same as the architects. For the collection of data, the individual experience of the author in the subway design,the information collected from the construction company through informal interviews, objects, materials and equipment purchased from the project company, design decisions taken in the process, the collaborative studies with other disciplines in the project process were used. As a result, it is aimed to develop an implementation method for Algo-

rithm Aided BIM in order to minimize user-focused time consuming and labor-intensive tasks with investigation of the case study. With the help of algorithm, users reduce this tension in the project planning environment. However, the algorithm needs to be managed and implemented by different users. For this purpose, the breakpoints, application approach and method for the transition to the algorithm aided workflow in the project phase are examined in the case study. The case study is conducted by an empirical test method. Theoretical assumptions were developed as algorithm aided workflow and tested with case study. A particular workflow planning and implementation method were developed for Algorithm Aided BIM. Breakpoints in BIM workflow were determined with the number of repetitive tasks in the project process, introduced by Reseller’s (1998) complex model definition and the list of repetitive tasks were expanded based on the author’s observations in relation with the the representation levels of BIM and the number of disciplines involved in the project process. In the beginning of the case study, , the intensive workforce in the project planning phase and time-consuming workflow are defined as breakpoints. Further to the definition of the initial breakpoints, involvement of each new breakpoint and their impact on overall efficiency were evaluated. Assessment of the subway project processes were considered as preliminary, final, and application respectively. The project processes were evaluated according to the 6 factors that can cause the breakpoints, which lead to transition to algorithm aided workflow. These factors were examined in terms of LOD, nD, maturity level of the model, number of disciplines, complexity of the project and number of repetitions in the project process. Based on the evaluation results, it has been assumed that either repetitive, iterative, recursive operations or the design model need to be reconstructed through algorithms where a tension was emerged causing the need for a breakpoint. Algorithm Aided BIM workflow were prepared in Dynamo software. However, it was not

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

Figure 3. Factors pushing users in subway design processes in BIM environment.

considered sufficient to use merely algorithms, provide optimization in the project process. Apart from those,there had been a necessity to integrate the Algorithm Aided BIM for the usage of different stakeholders. This is because the standards were generated to be open to any use in the selected construction company. When the preliminary project phase has been evaluated based on the above mentioned Figure 3 according to 6 breakpoints, the decisions in the earlier phase of the project were limited. Therefore, it was difficult to capture similar patterns according to design decisions. Furthermore there were few breakpoints in the modeling process that push the user to initiate new workflows. Algorithm aided workflow was not needed at the preliminary stages in the subway project because model components were consisting of low level of detail. Along with the increase of the data load in the project process, the need for algorithm aided workflow was started.

The steps at the final project were evaluated according to the 6 breakpoints as well. The final project phase was depended on collaborative works by users from various disciplines. Moreover, the level of detail of the objects was complicated, so the databased was enlarged. In the modeling section, the objects were encoded in the visual programming interface. The coarse layout plans have been differentiated after the inclusion of the electromechanical team in terms of equipment layouts, equipment coding, room entrances and exits, door directions and locations. Therefore, as the number of decisions taken in the project increased, the data and workload in the project have been also increased. To update the database, each object in this phase has been used with their own code, which were modeled in the relevant standard. However, there have been a lot of tasks that were repeating in the final project phase with new design decisions. There has been a need for controlling the data during ITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

repetitive operations. Otherwise, since it was not a geometric representation, mistakes were observed in the uncontrolled workflow. As a result, the number of breakpoints, the number of disciplines working on the project are causing serious workload and waste of time, which lead to adherence to algorithmic workflow at this stage. Furthermore, in order to be able to draw the project, it was necessary to translate BIM programs in accordance with the workflow and parameters. With 4D modeling, controlling different parameter and validation of the data were required as well as completing on time. For an efficient project process, regular checks and adjustments needed to be done.. However, as the project process has become more complex, it became more difficult to control. In the cases when controlling the code and the model became difficult, the user has searched new workflow in modelling process. This led to breakpoints in the workflow. Further to the evaluation of the application phase based on the specified 6 breakpoints, it has been observed that the detail levels of the objects were increased. Identification of the companies and writing these information into the objects and information complexity increased workload on the user. As the cost estimates and analyzes were made with 5D modeling in the application phase, the accuracy of the information in the database has also played an important role Although at the maturity level of 2, the data flow between different users appeared to be more efficient and easier, it took time and effort for a few stakeholders to learn and understand basics of BIM and algorithmic definitions. Coordination was difficult with project stakeholders in the further phases where shared data got more details. Moreover, the increase in data load have brought additional need for changes in the project and also accelerate the workload through repetition of the similar jobs. The more the complexity of the model has been increased in the application phase, the more control need has been occurred in each change. However, in large infrastructure projects involving thousands of entities, it has been very

difficult to control the model in each change. Parametric workflows were needed to overcome the difficulties regarding the control problem in BIM program. In a broader sense, when the application phase was evaluated, it has been difficult to transfer the system details of the identified companies and review the model, control the data load and process the information in the database. Each discipline worked in more detail at this stage. As a result, when the application phase was evaluated, it was the phase where stakeholders have needed the most algorithm aided workflow. In addition, improvements in the project process and software cause less labor and time spent in the project process of BIM tools. In the project development phases of the subway project, it has been observed that an integrated environment was needed in which Algorithm Aided BIM processes can be implied and executed. In order to reduce the heavy workload of the project process, not only algorithms but also standardisation for the stakeholders were needed to work in cooperative environment. The validity of the standards has been tested in BIM project processes. According to the result of the analysis, algorithms should be standardized in an environment that is open to business cooperation. Otherwise, discontinuities in communication among different users have been observed. In order to increase the adaptation of users with knowledge of different programs, clarity of algorithm needed to be improved by informing and separating function groups. In this way, the generated algorithm relatively has become more efficient and more open platform for articulation. Inferences from the subway project process and the platform, on which the algorithm was built, should be split into two parts: information and functions. As shown in Figure 4, information sharing and function sections of the algorithm are used to express company wide workflow clearly and spread the usage of the algorithm between different users. Information section: â&#x20AC;˘ The name of the person who prepares the algorithm,

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

Figure 4. Companywide usage standards.

• The version prepared by the algorithm, • If additional package is used, • A short text in which the algorithm is created, • If there is a file exported from outside, the extension of this file, • If the prepared algorithm is specially prepared for a project, • An informative article explaining how to solve the mistakes in the use of the algorithm, • If the existing algorithm has been developed, a summary of who has developed and the new logic should be written. Whenever there has been a change in the algorithm, other users automatically were informed about the about who made the change. In this case, the version of the Dynamo became important. This is because, there were different applications in the different version of the software which has caused confusions. When an error has occurred in the code execution, the version needed to be changed, users were aware that there might be changes in the blocks. Moreover, in the cases that additional package was used, the relevant information has been updated in the infornatiıın section so that other users could learn what additional packages to install before running the code. In this way, the disruption caused by the fact that the package was not in that user has been removed. It has been necessary to write a brief note for informing the logic of the algorithm. The changes might be forgotten even by the creator of the algorithm, when it is not used in the long run. The observations derived from the case study shows that to avoid reestablishment of the algorithm resulting from the loss of information in the process, a brief description and an approach to the problem should be clearly stated. Some algorithms are specially defined to the specific project or specific problem. It is necessary to write the infor-

mation about which project and when these algorithms are not used in the whole company. Users with different levels of programming knowledge are able to solve specific problems without the need for advanced users, as they are involved in the process. For this purpose, if the solutions of the mistakes is informed, users are solved without help of advanced user who has high programming knowledge. As algorithms are used in the process, they are modified or improved due to changes in new requirements or project decisions. If the summarizer of the algorithm and the summary of the innovations are included in the information section of the algorithm, the information loss between the development process is also prevented. As a result, Errors and reference for external file path in the information section is seen as necessity to decrease dependency of users to each other, when the project phase samples are tested from the logic of classifying algorithms, the idea of design patterns of Alexander (1977), the grouping methods used in Grasshopper visual program software and the standards prepared by White Research Lab’s Dynamo software. For this purpose, it is necessary to explain what the user should do in response to the error in parameters. In this way, the need to users who know advanced programming could be reduced. Thus, the ability of the users to generate a solution against a certain level of error can reduce the pauses in the process. For this purpose, as the project principle is guided, the solution should be expressed clearly in the guidance section as the error arises. Specifying the source of the file to be exported is to ensure that the relevant file is available without consulting each time to the people who have contributed in the process of preparing the algorithm. With this type of comprehensive information writing, stakeholders are trying to be adhered to new business steps by ITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

reducing the need of users with different levels of programming knowledge to other users. Function section: In the part where the algorithm is established, grouping and coloring of modules in according to the aims are effective approach as seen in case study. When user looks at algorithms, they could be read easily. As shown in figure 5.39 above, although inside of algorithm groups is not understood by users, purpose of groups could be understood clearly. The coloring and naming of the groups increase clarity to convey the purpose. In particular, explaining what to choose in the selection blocks makes it easier to adapt algorithm. As a result, when the samples at the projecting stage are tested according to the criteria of Woodbury’s (2010) classification algorithm, Alexander’s (1977) idea of design patterns, the grouping methods used in Grasshopper visual program software and the standards prepared by White Research Lab’s Dynamo software, the presence of the block at WIP is considered unnecessary and confusing. The code that is studied may be inaccurate or incorrect. It will be more accurate to work in a new folder instead of going through the same code. With the division of the function groups into interface and back sides, users with the low level of programming knowledge level is separated from the advanced user. In this way, different stakeholders can attend the process without breaking the code of logic. In addition, with the grouping of blocks in the interface, users with a low level of programming knowledge can more easily understand and contribute to the algorithm. As a result, it is observed that the workload can be reduced by algorithm aided approaches. Working on the joint project by different stakeholders in BIM shows that planning new workflow is also necessary. Otherwise, the new workflow move more independently in BIM process and there may be breaks in the project phases. Standards must be established so that new workflow can be used by other stakeholders. In order to test the validity of the prepared standards, the project process is still being investigated

through the ongoing subway project during the period of the survey. As a result of the analysis, it is concluded that algorithms should be created in the new workflow to reduce the tension arising from the heavy workload in the projecting process. Moreover, algorithm should be standardized to increase the adaptation of different stakeholders in the company like BIM processes. In the absence of standards, in spite of being able to optimize, it is monopolized by the limited user and it is difficult for new users to understand and adapt because of lack of standards. Algorithms are used to reduce stress due to heavy workload and time loss during the projecting phase, and the use of a common language within the company reduces the dependence of users on each other and reduces the pauses in the process. It is prepared for example of project offices that want to pass on algorithm aided workflow with the created approach method. 5. Conclusion The impact of using algorithm-aided approaches in the BIM environment were investigated through a case study with a particular focus on efficiency in project design and development process. • Algorithm aided approaches appear to reduce the time and effort of the user in the project development process. However, it has been observed that algorithm aided workflow conducted with BIM might cause a limited amount of stakeholders to become more dominant in the process than others which contradicts with the very basic idea of collaboration. Asynchronous access to the shared BIM model or accessing but not understanding the repetitive and recursive operations defined via visual algorithms might create gaps between the designers and the BIM model. • It is necessary to reduce the stress caused by the heavy workload in the project process and spread the companywide standards from the very beginning of project design in order to increase efficiency. Otherwise, the efficiency is provided either for the limited user, or it de-

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

pendents on the users who have advanced programming knowledge. • There is a need for users with advanced level of programming knowledge to establish new project steps that come with algorithm aided approaches. Otherwise, learning both programming and building a system can be more difficult and challenging than traditional approaches. This study aims to present highlights for the companies that intend to implement algorithm aided approaches in BIM during project design, development, management and construction processes. In terms of switching the core database and workflow from CAD to BIM, and/or from BIM to computational BIM, one of the most common problem for companies is the lack of standards. With this in mind, in the scope of the study an algorithm based model was proposed to be adapted and used in different contexts by various companies. The validity of the proposed model was tested through a subway project. The subway project was selected as a case study subject due to its containing repetitive operations and its level of complexity, the increasing need for coordination between different users and the frequent changes. Due to stakeholder’s contribution to the project in different phases and dominance of some actors to others, the need for spreading standards across the company is a necessity, regardless of infrastructure, superstructure project differences. It is shown that companies that use BIM programs will be able to pass on an efficient business process, and after the detection of the pattern that causes unnecessary and repetitive work in the project planning process, algorithm aided approaches and company-wide standards are expected to reduce the load on the user. Standardisation of algorithm aided workflow to have potential improve the adaptation and motivation of the users to new business steps. According to BIM manager of company, algorithms in standardisation prepared in the subway project become more understandable with the use of some of them. Acceleration of the process in the pro-

jecting process with the participation of other users and the decreasing dependency of the users to each other in the informing part and grouping of the algorithm and decreasing time loss for obtaining information in the process is seen. It is determined that the quality of the model increases with the decrease of error rates with more controlled project process. As a result, it has been observed that the spread of algorithm aided approaches throughout the company has prevented the loss of time and unnecessary workloads. Integration of the algorithms into the project phases is not a surprise that automation will be achieved in the process, but with the standards established within the company, time and unnecessary works are reduced, so cost and repetitive tasks are reduced and more comprehensive automation and efficiency are provided in terms of more controlled progress. Similar to the algorithm aided approach of BIM working environment, the cooperative environment for multiple stakeholders is important in terms of putting them through the metro project as much as possible in such integrated systems. References Aish, R. (2013). First build your tools, in Peters, BP and Peters, TP (eds) 2013, Inside Smartgeometry: Expanding the Architectural Possibilities of Computational Design, John Wiley & Sons, Ltd, Printer Trento Srl, pp. 36-49. Alexander, C., Ishikawa, S., & Silverstein, M. (1977). A Pattern language. Oxford University Press. Boeykens, S. (2012). Bridging building information modeling and parametric design. eWork and eBusiness in Architecture, Engineering and Construction: ECPPM 2012, pp. 453. Crafai. (n.d.). The maturity of visual programming. Acess Date: 02 November 2015, http://www.craft.ai/blog/ the-maturity-of-visualprogramming/. Eastman, C., Teichholz, P., Sacks, R., & Liston, K. (2011). BIM Handbook A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers, and Contractors (2nd ed., pp. 648). John Wiley and Sons. Humppi, H., & Österlund, T. (2016). ITU A|Z • Vol 15 No 1 • March 2018 • Ş. Yıldırım

Algorithm-Aided BIM. In Proceedings of the 34th eCAADe Conference, Oulu, Finland. Janssen, P., Chen, K. W., & Mohanty, A. (2016). Automated Generation of BIM Models. In Proceedings of the 34th eCAADe Conference, Oulu, Finland. Ondejcik, V. (2016). Dynamo Graphic Standards at White arkitekter AB. Access Date 30 April 2017,http:// dynamobim.org/dynamo-graphic-standards-atwhite-arkitekter-ab/ Rescher, N. (1998). Complexity: a

philosophical overview. Transaction Publishers. New Brunswick, NJ. Woodbury, R., Aish, R. and Kilian, A. (2007). Some Patterns for Parametric Modeling. in: Lilley, B. and Phillip, B. eds. 27th Annual Conference of the Association for Computer Aided Design in Architecture, Dalhousie University, Halifax, pp. 222-229. Woodbury, R. F. (2010). Elements of Parametric Design. Routledge, Abingdon.

Investigation on algorithm aided BIM standards to increase collaboration and optimisation in project phase: A case study

ITU A|Z • Vol 15 No 1 • March 2018 • 79-92

The sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macroand micro-geometries for acoustic performance Dagmar REINHARDT dagmar.reinhardt@sydney.edu.au • School of Architecture, Design and Planning, The University of Sydney, Sydney, Australia

doi: 10.5505/itujfa.2018.87894

Received: October 2017 • Final Acceptance: February 2018

Abstract Sound performance plays a significant role in the experience of space. Theatre and performance spaces provide a context where acoustic and spatiotemporal characteristics can be informed through the controlled robotic fabrication of macro- and micro-geometries, by using mathematical principles as a driver for design variations and machine code. This essay discusses a short history of relationships between sound and geometry; from acoustic reflection methods (Kircher, 1673); to a theatre seating matrix (Saunders, 1790); to positioning of individual listeners and their specific acoustic environments (Russel 1883); to sound concentrations in spherical shapes (Cremer, 1982, Vercammen, 2012); to current strategies for acoustically performative scattering surfaces (Reinhardt, 2012, 2015). The essay further introduces empirical research into robotic prototypes that test the acoustic effects of complex architectural geometries, with a focus on robotic fabrication of macro geometries that change the colouration of sound; and micro-geometric surfaces that can be applied to improve acoustic performance by scattering. It presents a 6-axis fabrication process for acoustic scale prototypes, based on a range of mathematical equations that regulate physical properties of spatial surfaces and pattern details. Here, generative tools and robotic tooling processes are linked to the angle and cavity depth in a surface medium. The essay concludes with a discussion and an outline of future strategies for the acoustic performance in multi-talker work environments or daily life scenarios. Keywords Robotic fabrication, Theatre, Acoustic performance, Scattering surfaces, Sound concentration, Micro-geometries, Speech intelligibility.

1. Introduction Sound is an immersive experience. We encounter it as in many different forms, as noise, melodies, voices or song. Our sense of sound is the oldest, most primal mechanism capable of establishing a culture through narration. The acoustic character of spaces feeds into our perception and experience. In fact, architecture ‘echoes’ or reverberates with sound, often without being explicitly designed to do so. All spaces deliver acoustic events by reflecting sound-waves projected within. This occurs relative to spatial characteristics; such as to geometry and shape, structure, materials, and surface finishes, but is also further dependent on the sound source, speaker and listener positions. Acoustic performance depends on the quality and character of sound propagation as a function of both overall spatial volume and surface properties, which combine to affect the spoken word or sound: through the overall shape (or macro-geometry) of space, and the character of its surface finishes (or micro-geometry). Depending on the desired effect, these can be shaped to decrease or increase sound performance, by absorption or by reflection. Furthermore, scattering can be controlled to change the direction of a sound, thus creating unique listening experiences. In spaces for the Temporal Arts, such as theatres, concert halls or churches, the sound of space needs to be highly specified for performance (Figure 1). A general sonic character and specific aural qualities are intrinsically tied to choreographed activities and programs. And while most performances are amplified, theatres and concert halls tend to exhibit curved, spherical, ellipsoidal and often multiple intersecting complex geometries that support non-amplified performances by concentrating sound. These enforcing geometries are mirrored in the primary character of the space itself, or in secondary spatiotemporal characteristics such as seating arrangements or stage setting. Simulating and forecasting sound is difficult because many parameters interplay and accurate computational predictions of sound propagation in

space have relatively high (computational simulation) affordances. Hence, in architectural acoustics, scale models and prototypes are commonly used, which are preferable to virtual simulations since they monitor the physical acoustic phenomenon itself (Peters and Olesen, 2010). Alternatively, a variety of theoretical paradigms are available for modelling the behaviour of sound, including analytic, statistical and numerical methods; based on ray, wave, and particle propagation; in any or all of the domains of time, frequency, and space. A computational prediction of performance describes sound through mathematical models (such as Matlab equations as a point of departure), and acoustic simulations (Odeon or Ease) support optimization for performance through the negotiation of multiple criteria. Most significantly, common 3D modelling and scripting software (McNeel Rhino and GH Grasshopper) enable workflows and transitions between the different software environments. To this extent, robotic fabrication can be deployed to bridge between design scope and desired acoustic performance, and the manufacturing of surface treatments or modules for spherical or vaulting architectural spaces. While traditional methods of producing scale model prototypes require high levels of manual craftsmanship, six-axis robotic fabrication protocols enable precise, fast, variable, detailed, and economic manufacturing of curved surfaces and surface patterns. Moreover, robotically cut 1:10 scale model prototypes can seamlessly integrate geometrical descriptions, and thus expand the spectrum of available acoustic options. In other words, geometry presents a mathematical source-code for the acoustic design of architecture, which can be simultaneously and variably linked to design, structural simulation, acoustic analysis, and fabrication protocols, and consequently supports the non-amplified acoustic performance of spaces. In the following, this research introduces theatre spaces and derives strategies for shaping sound based on historical developments of relationships between ray/wave propagation

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

Figure 1. Complex spherical geometries of domes and vaults in Hagia Sophia, Istanbul (a). Pattern sequences in the Blue Mosque, entrance detail, Istanbul (b). Robotic fabrication protocols for micro (c, top) and macro-geometries (c, bottom). Images and diagrams by author.

and geometry. It presents the development of successive case studies that investigate robotic fabrication of micro- and macro-acoustic geometries and patterns. The research continues to discuss a framework in which robotic and computational design and analysis codes across different scales directly link to mathematical equations that determine acoustic behaviour. Hence, sonic events such as the directional reflection of sound, typical sound characters or ‘sound colourations’ becomes possible. The research concludes with a methodology and workflow proposition for future research to further support the design of acoustic performance of spaces beyond a context of the temporal arts. 2. Context of performance spaces and theatre acoustics Theatre frames the choreography of human expressions and cultural practices in four-dimensional space. In this dynamic setting, performance can be considered two-fold, as a performance of actors within space, and of space itself. Performances deploy a multiplicity of factors, such as the choreography of bodies and movement; size and program organisation, seating, visual lines, machining, lighting, and most importantly the geometry of space. Classical theatrical performances involve a seated audience in a directional visual line, a space split between audience and a centre stage framed by

a proscenium arch (a picture frame). Its acoustic and visual operations are conditioned by a directional view, fixed seating, defined stage presentation. Usually, an amplified sound system with a centralised source distributes sound in space, as opposed to ‘naturally’ amplified sound through spatial conditions. In recent years, contemporary performances have shown a tendency to explore multi-directional spaces with potentially simultaneous presentations (Figure 1). In these events, a walking audience engages with performances from many different angles. In experimental theatres where ‘choreography elicits action upon action’ (Forsythe 2011), parallel events are proceeding simultaneously. The multiple sources, viewpoints and audience positions in such synchronous storytelling consequently require a different treatment of the environment and sound technology. If the strategy is not to bundle numerous sources into a centralised sound distribution and thus amplified performance, in which way can various discrete, local aural conditions be established? 2.1. Theatres In theatre, the very nature of performance relates to a spectrum of acoustic events. The capacity of an actor to speak into space, and the ability of an audience to understand the words or lines delivered is essential to the ‘soundscape’ (Figure 2). Furthermore, actors

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

rely on the acoustics of space to listen to their voice reflected back (commonly referred to as speech intelligibility), and in fact continuously adapt their voice projecting into space. This vocal impact is based on a multiple-criteria set; such as the length and speed of sound waves; the surface characteristics they hit; the angle in which they arrive and thus exit as sound reflection, amongst many others. Acoustic performance is mostly conditioned relationships between the sound source, and different scales of geometry that inflects the sound. Natural geometries that amplify sound can be found in Greek and Roman theatres and are the result of a curved geometry that frames the ‘theatron’ (θέατρον, Greek: a ‘space for watching’), a technical term for an evenly rising area of audiences (Pelletier, 2006). The early Hellenistic theatres used negligible backdrops and a staging area so that optimal sightlines were given. A flat area for the chorus (orchestra) is extended by a raised area for actors, with seating arranged in a segment of a circle, or semi-circular or semi-oval layout around the stage. In this bowl-like arrangement, both steps and risers reflect the sound and so increase the performance capacities of an open-air space (Figure 3, a). Further developments in Roman theatres shifted the actors from the orchestra, and onto the stage. This included introducing an expanded back-of-house and adding a backdrop that also served sound reflective functions. Rendering the stage as an independent spatial or structural element inevitably led to a separation of actors-space and audience-space. Most significantly, this directionality resulted in directional sound relationships, effectively limiting a previous omnidirectional experience to a 90-degree acoustic and visual field. Saunders was the first to present in ‘Treatise on Theatre’ a comprehensive pattern language for theatre layouts upon which a significant number would continue to be built (Saunders, 1790). Based on an actor as a central focus for both viewing and listening, he discussed oval, semicircle, quadrangle, to horseshoe geometries as potential space diagrams. Potential visual dis-

Figure 2. Temporal Arts in acoustic and visual performance: human voice and song (left: Eddie Perfect, Misanthropology), and choreographed dance movements (right: William Forsythe, Endless House).

Figure 3. Geometrical relationships between acoustic and visual performance in the Temporal Arts: Common Shapes for Roman and Greek theatre spaces (a); G Saunders’s diagrams of relationships between theatre geometries and sightlines (b), Saunders, Treatise on Theatre, 1790); and John Scott Russel’s ‘isacoustic curve’ (c). Russel, Edinburgh New Philosophical Journal 27 (1838).

ruptions caused by geometry are indicated, such as distortion of images with an increase over 45 degrees in viewing angle. Amongst primarily visual dominated options, some diagrams also refer to sound travelling, reflected by wall surfaces in a box, as opposed to within ellipsoidal geometries (Figure 3, b 6-7). The latter is significant since it represents an acoustic phenomenon that is mostly ignored until much later. Russell extended this two-dimensional catalogue of spatial and predominantly visual based options towards sound (Russell, 1838). His ‘isacoustic curve’ describes diagrammatically the way in which sound waves travel congruently with sightlines. Consequently, spectators should sit along the lines of equal sound intensity. By aligning head and shoulders of individual listeners according to distance and angle, a curve emerges whereby the steepness determines the optimum acoustic experience for everyone, with

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

a direct sound path (Forsyth, 1985). Significant is here that this establishes a three-dimensional sonic space, with an upward slope from front to back, virtually eliminating a broad back wall where reverberation could cause problems. Instead, each seat is a serial module that becomes part of a formation of sound reflectors and contributes to the overall dynamic in the area of performance. In short, geometries of space, stage setting, and acoustic and visual performance are intricately linked. The morphology and curvature of surfaces and the conditions within the surfaces themselves can generate widely diverse sound effects. 2.2. Sound as a function of space (macro-acoustics) When sound encounters a surface it can be absorbed, reflected, diffracted or transmitted, and will, depending on the morphology of said surface, result in different acoustic responses regarding sound reflection. Blancanus discussed acoustic phenomena in relationship to curved geometries (Figure 4) in ‘Sphaera Mundi - Echometria’ (Blancanus, 1620). Kircher traced reflections that focus in and are amplified by concave and convex surfaces, and thus laid the foundation for a concept of geometrical acoustics. In his compelling study ‘Phonurgia Nova,’ Kircher extended this theory into applied geometries, particularly with principles that differentiate sound phenomena

in arches, vault, and domes (Kircher, 1650). Detailed studies show sound rays in a circle that originates from one centre point are reflected by an opposite concave surface, and refocused within the same centre point, thereby creating ‘white noise.’ In contrast, ellipsoid geometries produce sound reflections that continue towards an opposite centre point along the same axis, and so avoid sound concentration. Huygens introduced his wave theory which explicitly models acoustic behaviour following principles of optical geometries (Huygens, 1690). Sound waves are treated similar to light rays, and so demonstrate wave propagation. The ray entrance angle reflecting from a surface determines an equal exit angle relative to the surface normal. Furthermore, the angle is naturally domesticated by the overall surface geometry of the wall. Consequently, geometry can be understood as a tool with which to model sound behaviour - as sound trajectory across space, or in an incremental development in volume/density. Langhans detailed such incremental sound build-ups that form successive wavefronts through secondary sound reflections in his treatise on acoustics, ‘Ueber Theater oder Bemerkungen über Katakustik in Beziehung auf Theater’ (Langhans, 1810). In spaces with a circular ground plane (with a centre based sound source), sound rays meet the wall at the same time, are reflected, and a sound wave and subsequent wavelets that build up a wavefront, and will converge simultaneously upon the source position. Reflections of second order that render audio concentration in the audience area that negatively impacts on the speech quality. Langhans proposed secondary sound measurements to overcome these unwanted echoes such as the use of ellipsoids. When sound traverses a surface, it potentially enables private conversations between two parties. Surfaces curved in two directions such as ellipsoids result in a redirection in the opposite centre, as a result of this deflecting sound concentration. These significant observations, unfortunately, remained as such for a long time. As the research argues, these historical precedents can be strategically

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

Figure 5. Kircher (1673) introduces entrance-exit angle relations in ray reflections for concave surfaces (a), and a series of tectonic plates that direct sound pathways (b), Kircher, Phonurgia Nova. C Langhans (1810) further describes a patterning strategy for surfaces describes to produce scattering effects (breaking successive sound wave-fronts in circular spaces (c, top, plate 31,32); sound pathways in a human ear (c, bottom, plate 50,51, both in: Ueber Theater oder Bemerkungen ueber Katakustik in Beziehung auf Theater).

employed to shape relationships between sound and space. By informing the macro-structure of space, its circular and ellipsoidal geometries, sound reflections can be intensified. This presents the first strategy for ‘analog’ amplification whereby the geometry has a direct impact on the behaviour of the overall ‘soundscape’. 2.3. Sound as a function of surfaces (micro-acoustics) An overall geometry of space can inform sound behaviour, but so can micro characteristics, patterns and surface conditions. The scattering of sound, when reflected by a surface, can be a desirable characteristic in the acoustic treatment of spaces, and is achieved through patterns (Cox and D’Antonio, 2009). Kircher’s early descriptions of speculative acoustic experiments present such a spatial positioning of planar sound reflectors (Kircher, 1673). Some angled panels modify the properties of the space; instead of traversing the concave wall, the sound is scattered by the incidence angle and reflected back to the source origin that coincides with the audience position (Figure 5,b). The reflections decrease into a series of syllables (Latin.; clamore-amoremore-ore-re), that successively changes the meaning. This demonstrates a compelling variation to the common sound phenomenon of wide reflections of sound (scattering), which represents one of several major acoustic strategies

(amongst sound absorption, reflection, diffraction). As Langhans also documented in ‘Ueber Theater’, scattering principles work across different scales (note here the scale-shift from wall geometry to passage into a human ear). This can be achieved through a variety of acoustic patterning strategies for surfaces that produce scattering effects, both through concave and convex subtractions, whereby sound is reflected irregularly over a wide range of directions (Langhans, 1810, Figure 5, c). Consequently, geometry can be understood as a tool with which to model sound behaviour as implemented through discrete or localised pattern topographies. Whereas architecture seldom implements considerations of sound relative to spatial form or shape, in the field of audio-acoustics these observations are fundamental. In general, terms, scattering results from variation in the physical surface such as curvature, relief forms or textures, and changes in contrasts in material acoustic properties (Sabine 1964). Specifically, varying the angles and depths of a surface with relatively shallow surface modulation achieves scattering at high frequencies, such as the spoken word (Reinhardt et al., 2014, 2016, 2017). For high-frequency scattering, aperiodic patterns are preferred so as not to reduce the complexity of the reflected sound field (Bonwetsch et al., 2008). Additionally, a known factor is that surface rough-

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

ness increases sound absorption. And although scattering performance is relatively hard to predict, incremental variations in of pattern geometries can be computationally controlled via depth and width. Consequently, a second strategy for shaping sound can be derived: the micro-structure of surfaces, its rhythms, densities and depths, impacts on sound reflections. This presents another potential for ‘analog’ amplification whereby the geometry has a direct impact on local modulations of sound effects. We defined these strategies as a potential for the development of sound prototypes, with accurate propositions for an explicit listening experience. For the geometrical, physical and mathematical modelling of the sound field of a source in space, we used mathematical models as source code (a DNA), and reverse engineered it for generations of prototypes. 3. Robotic prototypes for shaping sound Based on the two strategies for micro- and macro acoustic sound shaping, the research explored a series of robotic case studies that extend scripted codes or codes derived from signal processing software directly to the 6axis robotic manufacturing for non-standard spaces and non-periodic patterns (Reinhardt et al., 2017). Previous robotic research developed between 2013-2017 is presented here in a comprehensive discussion for the first time. We discuss development of sound performance - from the robotic fabrication of macro geometries that change the colouration of sound to micro-geometric surfaces that can be applied to improve acoustic performance by scattering. Several 6-axis fabrication processes for acoustic scale prototypes are discussed, based on a range of mathematical equations that regulate physical properties of spatial surfaces and pattern details. This approach thus expands the scope of geometric surfaces by integrating scripting logic, surface angles and depth, and toolpath, thus enabling successive acoustic design variations that were tested for proficiency. This is undertaken to identify a framework and pathways

for architecture and acoustics towards shaping non-amplified acoustic performances of spaces, towards applications in multi-talker work environments or daily life scenarios. In the following, the empirical research studies share the interfacing of parametric design, structural analysis, acoustic analysis and the potential of robotic fabrication. For all case studies, the following segments were shared to derive threshold criteria: • specification of the architectural design parameters in conjunction with acoustic aims (e.g., colouration or scattering coefficient spectrum); • scripting codes or mathematical sequence for specific geometries, by computational modelling and scripting software (McNeel Rhino, GH Grasshopper, Matlab); • evaluation in acoustic simulation (ODEON, Ease); • simulated for robotic fabrication (KUKA|prc), and fabrication of physical scale model test samples (scaled prototypes or discs); • and acoustic measurement and analysis of sample or prototype performance; and • further design iteration and refinement. The complexity of acoustic reflections required a workflow rationale, and hence this stream allowed extending from scripting surfaces towards the physical measurement of scale prototypes as an essential part of the design and validation process. 3.1. TriVoc | Robotic subtractive cutting of macro-geometries for a harmonic trichord In the first prototype ‘TriVoc’, the research investigated in which way acoustic parameters that inform or consolidate complex geometries could be embedded. This required operations between different computational packages of simulation software (Comsol Multiphysics), plugin (McNeel Rhino/ Pachyderm), and mathematical equations (Matlab). Specifically, the generic shape (ellipsoidal, spherical or dome structures) were altered by adapting the height, dimension and centre point. The robotic case study ‘TriVoc’ (Reinhardt et al. 2013, 2014) uses this

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

Figure 6. Shaping macro-geometries through robotic fabrication. Robotic acoustic prototype 1:4 scale model (left). Concept development (right): Generating Matlab function for primary geometry between homogenous 500-500-500 Hz space (as) and tuned ellipsoids as harmonic 400- 500-600Hz space(b) for ellipsoids; RhinoPython Code (c) and mapping producible space dimensions with constraints of robotic work envelope (f). Diagrams by author.

to explore the harmonic inflexion of human voices through customised curved macro-geometries of spatial surfaces (Figure 6). The basic geometric/mathematical prompt was derived from Matlab (a computational and signal processing environment widely used in acoustics) which generates x/y/z coordinates for the ellipsoids. Based on the input parameters of distance between foci, and the respective tuning frequencies of each ellipsoid, this function was deployed towards two scenarios: three identical intersecting ellipsoids, and the more complex scenario of different three intersecting ellipsoids. By tuning the physical geometry of each ellipsoid, such as varying the distance between foci, the three intersecting ellipsoids were ‘tuned’ (in Hz). Thus, some geometrical solutions could be derived that look and sound very different to each other. This produces a three-person conversation space changes the tonal character or ‘coloration’ of voices and produces a ‘triad,’ or three-pitch chord. These base conditions were then integrated into machine code (RhinoPython/GH), and mapped onto robotic manufacturing codes (kuka|prc) so that the design immediately responds to fabrication constraints, and a capacity of actual (acoustic) performance (Comsol Multiphysics). A functioning (acoustically proficient) prototype was produced at 1:4 scale through robotic milling, which required scaling in

shape, and impacted on sound, but fundamentally maintained the most critical parameters of control over robotic toolpath: control over curvature and availability of resulting pitch for physical analysis. The intersection between Matlab, RhinoPython/ GH enabled control over the sound pitch and spatial design, to adapt the desired ‘soundspace’ continuously and repeatedly. 3.2. RoboFlowl | Robotic milling of micro-geometries for isocurves As opposed to the previously described study, the second, ‘RoboFlowl’, focused on the surface condition of micro-patterns to inform acoustic performance (Reinhardt, 2014). As scripted geometry base, two variations were generated in GH Grasshopper (a plug-into McNeel Rhino/visual scripting environment); with a hexagonal periodic and hexagonal deformed pattern; and a vector-based pattern adopting a customized script ‘Flowl’ (Figure 7). Both developed zones of highly differentiated depth across the surface, with individual facets varying in depth, height, directionality, resulting in sound diffusive properties of test discs that acted as acoustic pattern prototypes. These 1:10 scale model discs were tested by random incidence measurement in a scale-model reverberant room, placed on a turntable, and synchronously averaged impulse responses are obtained for different source and receiver positions from

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

Figure 7. Micropatterns. Shaping micro-geometries through robotic fabrication. Robotic acoustic prototype 1:1 scale model (left). Concept development (right): GH protocols for change of reflective area in scattering disc building zones around attractor points and changes in vortex field (a); specular reflections with sound rays reflected by an angled surface (b). Robotic fabrication syncing toolpath and angle of effective surface (c); and 364 facets that produce a soundfield (d, e). Images and diagrams by author.

the material sample. From these, the scattering coefficient was calculated (the ratio of acoustic energy reflected in a non-specular manner to the total reflected acoustic energy). Significant results were yielded for scattering at frequencies above 1 kHz from the ‘Flowl’ surface, due to depth and number of ridges/valley volumes, which thus proofed an effective strategy for scattering surfaces on a micro-basis for human speech performances. The robotic fabrication for the RoboFlowl prototype used the advantage of a relatively simple geometrical rule for deforming a collection of individual lines relative to one attractor and its adjacent neighbours. Every single line is a spline, but can be directly linked to the robotic toolpath, with the angle of the milling tool predefined and variable along the curve, resulting in the depth of valley that must be achieved to provide scattering. Instead of a workflow with multiple passes along splines, a pair can produce the surface void, and thus produces the acoustic scattering effect. This rule set interfaces directly with the robotic parameters, which include here toolpath and defined an angle of the milling tool; multiple passes along isocurve; the distance of the robotic end-effector to a material surface, and depth and surface angles of voids as a variable along the isocurve. In this manner, the

approach concatenates computational design, acoustic analysis and robotic fabrication, which expands the potential scope of micro-geometric surfaces by integrating scripting logic, surface angles, and depth, and toolpath, thus enabling successive acoustic design variations that can be tested for proficiency. Limitations resulted from the fact that out of over 300 some 34 still had to be manually adjusted. However, the 1:10 prototypes could be scaled up to 1:4 to develop acoustic ‘regions’ within the geometrical envelope, and so act as proof of concept. 3.3. PRotoDNA | Sequenced Milling of Micro-geometries In the third study series, single robotic test samples were limited to fabrication and required manual adaptation of original design script and robotic manufacturing process. Consequently, transfers between pattern scripting and flexible robotic fabrication for acoustic scattering needed to be significantly expanded to arrive at an efficient production of multiple series of 1:10 scale model prototypes, so that statistical information on relations between surface area and mean depth for the scattering of high frequencies could be derived (Reinhardt 2017). Hence, the basic premise for ‘PRotoDN,’ are continued developments as direct robotic manipulations of micro-acoustic patterns

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

Figure 8. Micropatterns as DNA of pattern script for robotic fabrication. Flocking based script for robotic fabrication (left). KUKA\prc simulation with hotwire endeffector (a), sound discs relative to cutting profile changes (b) ; acoustic profiles for scattering and absorption performance (c), and reverse-engineering a mathematical relation between valley total, density and pattern depth. Images and diagrams by author.

(Figure 8). Since scattering is not the only acoustic surface characteristic that can benefit from high degrees of freedom digital fabrication, the research continued into developing a spectrum of micro-design of special patterns of reflection, highly tuned absorption, and potentially other unconventional acoustic surface behaviours. Robotic fabrication extends here randomness (or non-periodicity) through motion control, tool-path angle and shape of end-effector (wire element). Current computational design approaches apply chaotic or unpredictable surface patterns to derive ranges of aesthetic solutions based on similar premises. As a concept of variation, randomness can be part of a highly controlled design process, which can be of particular significance in the context of acoustic performance in architecture. Surfaces deliver better acoustic scattering performance when based on aperiodicity. Hence, randomness can contribute to a broader range of diversity in the architectural envelope, when acoustic performance is understood statistically. The introduction of a degree of random variation in the surface has the potential to increase the frequency range of useful scattering, or absorption of sound – depending on the depth of valleys, and diversity of the pattern. As an empirical study of the relationship between bespoke fabricated surfaces and acoustic scattering tested degrees of variation in 1:10 scale

model prototypes. Based on a single script, the robotic subtractive process was interrupted or partial lines of code in fabrication, executed. This enabled us to develop a pattern range, to derive statistical data on acoustic properties of these surfaces, and to deduce design rules. The robotic patterns resulted from variations in robotic trajectory, depth, and sequence were random, non-periodic and non-directional. Each 1:10 scale model was assessed with detailed measuring scattering coefficients and random incidence absorption coefficients of the discs, using a scale model reverberant room. As a result, robotic manufacturing through spline curves offered a much more significant scope for variability in the production of sound discs, due to an easily accessible transfer of GH code to robotic machine code; maximum depth of valley could be continuously achieved across the surface. Acoustic testing further confirmed that similar densities of valley patterns cause similar acoustic behaviour, with simple robotic deposition of patterns on the surfaces, with a periodic spatial design of surfaces produces a periodic scattering spectrum with regular, visible peaks. These results enable differentiation of ratio between surface area, depth of cut and pattern frequency. The research found that measurement and integration of surface area of the disc could be expressed as a ratio to the flat disc before

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

cutting. The research also found that descriptors for the root-mean-square (RMS) depth of the surface (subtracting mean depth) and the circular FFT power spectrum around the surface (at various radii) could be derived. In this manner, from the same code but with a change of tool, both scattering and absorption highly efficient surface could be produced through very few robotic tooling lines, and with a wire profile that generated an undercut. This research then results in an archive of different acoustic behaviours that can be engineered and orchestrated over a larger field, and thus deliver statistical data on surface effects for absorption, directional reflection, and scattering. 4. Discussion In these studies, the interdisciplinary relationships, exchange, and collaboration between previously separate disciplinary areas contribute to a multi-criteria performance (Figure 9). Across the three robotic prototypes, the research has explored several acoustic concepts and case studies for non-amplified space. As a result: The prototype series ‘TriVoc’ (2013/14) produced research into the macro-geometry of three intersecting ellipsoids, with the aim to produce a space that resonates with harmonic tonal effects. While this initial study proved effective, the research considered a larger setup of spatial geometries expansive due to milling toolpath, and so pursued further the original concept of further refining the overall space. The prototype series ‘RoboFlowl’ (2015/16) investigated initial research into general trends for robotic fabrication of micro-acoustic patterns, such as the relationship between the depth range of the relief in the surface pattern and the frequency range for high scattering values. This aimed at generating a first fundamental understanding of pattern tendencies between planes and splines, and testing potential of robotic fabrication. As a result, general trends were determined for acoustic proficiency, here relative to the direct and implicit connection between guiding the robotic toolpath angle (through milling), and the resulting effectiveness in sound reflection angle.

Prototype series ‘PRotoDNA-1’ (2016/17) further explored the toolpath as a direct prompt for acoustic effectiveness and further explored the potential of gradient patterns. While the aim of this first series had been to find an empirical relationship between a physical parameter and acoustic result, the research generated here algorithmically differentiated patterns that were further intuitively changed. This was achieved through variability in toolpath and the defined angle of the subtractive tool, through multiple passes along isocurve, a distance of end-effector to a material surface, and the depth and surface angles of valleys variable along the curve. Across 38 sample prototypes, the research manipulated a range of machine code by an interruption, recursion, or shift to another surface, so that parts of patterns in different locations were generated. As a result, the pattern generations within multiple criteria (parametric modelling to scale model production to physical simulation) expanded the archive or acoustically active surface patterns drastically. This enabled us to derive tendencies that needed to be further investigated. Prototype series ‘PRotoDNA-2’ (2016/17) thus reverse engineered the previous study, as it derived mathematical equations and a working protocol for gradients of pattern densities across a surface, resulting in a specific acoustic performance. This further approach was geared towards including a differentiation ratio between surface area, depth of cut and pattern frequency; through measurement and integration of the of surface area of the disc, which can be expressed as a ratio to the flat disc prior to cutting; the root-meansquare (RMS) depth of the surface (subtracting mean depth); and the circular FFT power spectrum around the surface (at various radii) as the most detailed approach. As a result, a controlled and large spectrum of scattering-effective acoustic micro-patterns could be achieved for robotic subtractive processes, so that new design strategies for an acoustic performance in the built environment can now be applied.

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

Figure 9. Workflow and range of variable parameter. Development of base geometries for spaces and intersections, principle testing of structural performance principle analysis of acoustic behavior of macro – and micro-geometries, application of robotic manufacturing protocols for modules or components, and surface patterns. Note that this is not a linear process, but recursive. Diagrams by author.

A typical protocol would need to expand from small-scale acoustic optimisations to the integrated structural performance within complex geometries on a scale of buildings (Figure 9). A workflow consequently could be initiated, to: • identify and establish primary geometries (a) in 3D modelling software (McNeelRhino) or acoustic geometries (Matlab), • Produce initial acoustic analysis to determine position of sound source in (b) significant volumes, or in micro-pattern sequences (c) (BK Odeon, EASE), • if using spatial volumes, establishing parametrization (Grasshopper GH) of structural system (d) rigid shell system or (e) structural beams to morph geometry (f) height, (g) diameter, and (h) shift of central sphere point (abstraction of spheres), and test structural capacities and performance (karamba) • apply a second acoustic analysis based on macro parametric variations comparing (i) ideal and (j) deflected centre points • apply third acoustic analysis based on micro parametric modifications, relative to audible effect (k) reflective, or (l) absorbent, or (m) scattering distributed across the surface

• simulate robotic fabrication of modules or components in KUKA|prc, relative to work envelope (n) and acoustic pattern generation (o) • produce robotic fabrication of modules for an envelope, or patterns (p), and • extend robotic fabrication under the support of sensing technology and head-torso simulation to further enhance or change acoustic performance (q). To be continued. Such a workflow would enable the strategic development capable of interfacing between disciplinary areas, and a framework for collaborations that prototype a spectrum of acoustic performance for real spaces, and future spatial simulations. 5. Conclusion Robotic fabrication presents a huge potential for fine-tuning accentuated patterns relative to high frequencies for a person or an audience situated at any point in (a performance) space. As the research has discussed, sound propagation can be engineered as a function of spatial volume and surface properties. By interfacing computational scripts, mathematical source codes, and structural analysis and acoustic analysis, and then fine-tuning parameter (of dimension, distance,

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

height, curvature, focus point, sound source and audience position), sonic events can be generated, controlled, fabricated, and further evaluated and continued. Consequently, a compelling and distinctive topography of multiple scales and dimensional geometries can be produced. A diversity of spaces that enhance acoustic performances/experiences by delivering a combination of beauty and sound reflection: sequential and multi-dimensional choreographed and differentiated soundscapes. Most significantly, this can be done without being dependent on the enclosing envelope, and through precise and controlled approach with local contexts. Future research will need to be undertaken for the application in multi-talker multi-speaker environments, where different and contrasting acoustic protocols are choreographed. This can then extend criteria, conceptual framework, and robotic fabrication processes to the acoustic surface treatment of existing surface geometries, or to the conditioning of complex curved surfaces that can be sound-active on a larger scale in architectural space(s). This will enable fundamental research into other areas where unamplified speech or performance takes place (such as multi-talker workspaces, community halls, commercial open zones, or public transport infrastructures). In that manner, and building upon the discussed strategies, sound amplification for audiences, and retro-reflection for open workspace scenarios, and sound scattering for audio ambience can improve acoustic performance across a broad spectrum of everyday spaces. Acknowledgements This research has been developed through interdisciplinary collaboration, for which the author acknowledges team members: Alexander Jung, Densil Cabrera, Luis Miranda, Gabriele Ullacco, Matthew Hunter, Dylan Wozniak-O-Connor, Rod Watt, Lynn Masuda, and the postgraduate students of the CodeToPro 2016 elective. The research received SEED funding (2015, 2016 and 2017) of the University of Sydney. Robotic fabrication by the Architecture Robotics Lab/DMAF, and

acoustic testing by the Audio & Acoustics Lab, both at the Sydney School of Architecture, Design and Planning, The University of Sydney. References Cremer, L. and Mueller, L. (1982). Principles and Applications of Room Acoustics. Applied Science, New York. Blesser, B. and Salter, L. (2007). Spaces Speak, Are You Listening? Experiencing Aural Architecture. Cambridge: MIT, 70. Bonwetsch, T., Baertschi, R. and Oesterle, S (2008). ‘Adding Performance Criteria to Digital Fabrication Room-Acoustical Information of Diffuse Respondent Panels’. Acadia Conference Proceedings, Non-Standard Production Techniques Tools, Techniques and Technologies-Adding Performance Criteria to Digital Fabrication, Minneapolis, Minnesota, 364– 369. Burry, J., Davis, D., Peters, B., Ayres, P., Klein, J., Pena de Leon, A. and Burry, M. (2011). ‘Modelling Hyperboloid Sound Scattering: The Challenge of Simulating, Fabricating and Measuring’. In Gengnagel, C, Kilian, A, Palz, N and Scheurer, F (eds), Modelling Symposium. Berlin: Springer-Verlag, 89–96. Cox, T. and D’Antonio, P. (2009). Acoustic Absorbers and Diffusers: Theory, Design, and Application. London: Taylor & Francis. Dalenbäck, B.I., Kleiner, M., and Svensson, P. (1993). ‘Audibility of Changes in Geometric Shape, Source Directivity, and Absorptive Treatment - Experiments in Auralization’. Journal of the Audio Engineering Society 41(11), 905-913. Forsythe, W. (1985), ‚Choreographic Objects’. Frankfurt: Forsythe Company. http://www.williamforsythe.de/ essay.html. Keating, S. and Oxman, N. (2012). ‘Robotic Immaterial Fabrication’. In Brell-Cokcan, S., Braumann, J.(eds) Rob|Arch 2012: Robotic Fabrication in Architecture, Art And Design. Wien, New York: Springer. Kircher, A (1620). Phonurgia Nova. Reprint (1970). Hildesheim: Olms. https://monoskop.org/File:Kircher_ Athanasius_Phonurgia_nova.pdf

Sound of space in 3 robotic prototypes: Introducing 6-axis robotic fabrication to shape macro- and micro-geometries for acoustic performance

Knudsen, V.O. (1932). Architectural Acoustics. New York: John Wiley Academy. Kolarevic, B (2005). Performative Architecture- Beyond Instrumentality. New York: Spon Press. Langhans, C F (1810). Ueber Theater oder Bemerkungen ueber Katakustik in Beziehung auf Theater. Berlin: Julius Eduard Hitzig. Pelletier, P (2006). Architecture in Words: Theatre, Language and the Sensuous Space of Architecture. London: Routledge. Peters, B: 2010, ‘Acoustic Performance as a Design Driver: Sound Simulation and Parametric Modeling using SmartGeometry,’ in International Journal of Architectural Computing, No 3, Vol 8. Peters, B. and Olesen, T. (2010). ‘ Integrating Sound Scattering Measurements in the Design of Complex Architectural Surfaces -Informing a parametric design strategy with acoustic measurements from rapid prototype scale models’. ECAADE 28 Proceedings. Rasmussen, S E (1964). Experiencing Architecture. Yale: MIT Press. Reinhardt, D, Cabrera, D and Hunter, M (2017). A Mathematical Model Linking Form and Material for Sound Scattering - design, robotic fabrication, and evaluation of sound scattering discs: relating surface form to acoustic performance. Çağdaş, G, Özkar, M, Gül, F, Gürer, E (eds.), Future Trajectories of Computation in Design, 17th International Conference, CAAD Futures 2017, Istanbul, Conference Proceedings, pp. 150-163. Reinhardt, D and Cabrera, D (2017). Randomness in Robotically Fabricated Micro-Acoustic Patterns. Janssen, P, Loh, Raonic, A Schnabel, MA (eds.). Protocols, Flows and Glitches, Proceedings of the 22nd International Conference of the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA) 2017, Hong Kong. Reinhardt, D, Cabrera, D, Jung, A (2016). Towards a Micro Design of

Acoustic Surfaces - Robotic Fabrication of Complex Pattern Geometries. Reinhardt, D, Burry, J, Saunders, R, (eds.) (2015), Robotic Fabrication in Architecture, Art and Design 2016, Springer International Publishing Switzerland, pp. 136-149. Reinhardt, D, Martens, W, Miranda, L (2012). Acoustic Consequences of Performative Structures - Modelling Dependencies between Spatial Formation and Acoustic Behaviour. Achten, H, Pavlicek, J, Jaroslav, H, Matejdan, D (eds.). Digital Physicality - Proceedings of the 30th eCAADe Conference - Volume 1 / ISBN 978-9-4912070-2-0, Czech Technical University in Prague, Faculty of Architecture (Czech Republic) 12-14 September 2012, pp. 577586.ISBN:978-9-4912070-3-7. Russell, R. (1838). ‘Treatise on Sightlines’. Edinburgh New Philosophical Journal 27. In Forsyth, M. (1985). Buildings for Music. Cambridge: MIT Press, 236. Saunders, G (1790). A Treatise on Theatre. London: Taylor. Sabine, W. C. (1964). Collected Papers on Acoustics. New York: Dover. Schroeder, M. R. (1979). ‘Binaural Dissimilarity And Optimum Ceilings for concert halls: More lateral sound diffusion’. Journal of the Acoustical Society of America 65(40), 958-963. Schroeder, M. R (1975). Diffuse sound reflection by maximum− length sequences. The Journal of the Acoustical Society of America, 57(1), 149-150. Williams, N., Davis, D., Peters, B., De Leon, AP., Burry, J., Burry, M. (2013). ‘FABPOD: an open design-tofabrication system’. In Stouffs, R., Janssen, P., Roudavski, S., Tunçer, B. (eds). Open Systems. CAADRIA, 251–260. Vercammen, M. (2012). Sound Concentration caused by Curved Surfaces. Eindhoven: University of Technology, Eindhoven,The Netherlands. Issue 163 Bouwstenen series, 4. Vorländer, M. (2008). Auralization: Fundamentals of Acoustics, Modeling, Simulation, Algorithms And Acoustic Virtual Reality. New York: Springer.

ITU A|Z • Vol 15 No 1 • March 2018 • D. Reinhardt

ITU A|Z • Vol 15 No 1 • March 2018 • 93-105

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture Güzden VARİNLİOĞLU1, Burkay PASİN2, Hugh David CLARKE3 1 guzdenv@gmail.com • Department of Architecture, Faculty of Fine Arts and Design, Izmir University of Economics, Izmir, Turkey 2 burkay.pasin@ieu.edu.tr • Department of Architecture, Faculty of Fine Arts and Design, Izmir University of Economics, Izmir, Turkey 3 hugh.clarke@ieu.edu.tr • Department of Architecture, Faculty of Fine Arts and Design, Izmir University of Economics, Izmir, Turkey

doi: 10.5505/itujfa.2018.72623

Received: October 2017 • Final Acceptance: November 2017

Abstract Theories and methods of integrating digital tools into the architectural curriculum cannot be conceptualized as simply the merging of computerized tools with conventional formulations of design. This paper focuses on a case study of a workshop entitled “Mission Mars 2024: A Biomimetic Structural Organism”, as part of the studio course ARCH 202 in the spring semester of 2017 at Izmir University of Economics. It explores the use of digital architectural design tools in the context of outer space architecture, and the use of biomimicry as a design approach. We encouraged students to explore various stages of Oxman’s digital design ontology at the design level, and to employ various CAD/CAM tools as well as Virtual Reality (VR) and 3D representation methods. It is important to emphasise that the curriculum is a studio-based education with limited access to additional technical classes. Part of our aim was to integrate this content into the studio and allow students to explore new methods of design development. In order to free the students from conventional architectural preoccupations, we particularly chose on the surface of Mars. The paper presents a critical approach to understanding the impact of digital tools and methods on the learning outcomes of the students, which are discussed and demonstrated based on four studio outcomes. Keywords Architectural curricula, Biomimicry, Digital tools, Integrated studio model, Space design.

1. Introduction and background Design educators are facing the challenge of integrating computational tools into architectural education. The ability of these tools to work at a parametric, algorithmic and representational level necessitates an update in traditional methods of architectural education (Kotnik, 2010). Since the adoption of CAD/CAM technologies by other industries during the 1970s (Corser, 2010), emerging technologies in design and fabrication became increasingly prominent in architectural education and practice, particularly during the last two decades. As a result, computational approaches in architectural design education (Oxman 2006, Oxman 2008; Oxman and Gu, 2015; Özkar, 2007) and the insertion of digital fabrication tools in architectural curricula (Celani 2012; Duarte et al., 2011; Gül and Simisic, 2014) have been investigated in many architecture schools from multiple points of view. In many cases, the availability of tools, methods and resources in digital design setups are flexible, and open to change. In traditional design schools however, with less emphasis on digital tools, the matter of how these skills are introduced is important. A more traditional view taken by some schools is that too much emphasis on digital tools may somehow lead to a disconnection with existential space during the design process. Therefore, spatial design in extra-terrestrial environments are considered premature at second year studio level. This paper challenges some of these traditional views and investigates how digital fabrication tools are changing the design process that the students traditionally taught. How can students merge the conventions of architectural regulations with unconventional design contexts using new digital tools? How can a biomimetic design approach be combined with digital tools? This paper describes a two-week digital fabrication workshop in which students were asked to use digital tools to design a biomimetic organism on Mars. As a teaching group, the project allowed us to investigate the extent to which digital design and fabrication tools contributed to students’ design

thinking. Considering that there are minimum architectural constraints and contextual references in such an extraterrestrial space, we aim to find out whether students can freely explore the digital tools to come up with unconventional, creative and innovative design solutions. To evaluate our approach for integrating digital thinking, application, and experience in architectural education, we follow a comparative case study method in which Computer Aided Manufacturing (CAM) as fabrication method, Virtual Reality (VR) as a representation tool and biomimicry as a design approach played an integral part of the two-week workshop. 2. Digital architectural design (DAD) Oxman (2008: 106) makes a distinction between Computer Aided Design (CAD) and Digital Architectural Design (DAD). CAD methods, its principles and theory utilise paper-based design methods, whereas digital architectural design methods suggest new types of a form, grammar and concept. This reflects a shift in conventional architectural design processes because rather than being just a tool, digital architectural design methods have the potential to become the whole process itself. Principles, theories, and methods of CAD were based on imitating paper-based approaches. The new relationship between digital form and digital processes have contributed today to the emergence of a new conceptual vocabulary and domain knowledge which has led to a paradigm shift in design. It is important to note that Oxman (2006: 39) indicated the emergence of a new ideology by emphasising the conceptual conflicts between traditional and digital design. Oxman (2008) proposed five paradigmatic classes of digital design models, based on the various relationships between the designer, the conceptual content, the design process applied, and the design object itself: • CAD models: descriptive by employing various geometrical modelling and rendering software, but have little qualitative effect on design thinking, and are essentially isomorphic with paper-based design models (representational computability).

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

• Formation models: structured geometric or formal digital processes providing designers with a high level of digital interaction and control, a threshold between digital/ non-digital models • Generative models: a computational mechanism for formalized generation process. • Performance models: a process of formation or generation that is driven by a desired performance. • Integrated compound models: a complex mixture of the above. Oxman’s ontology emphasises the abandonment of typological and deterministic approaches in favour of the creation of generative and performance models. Still, at the representation and manufacturing level of computing, digital tools are widely used. Affordable CAM and low-budget VR devices are reshaping architectural curricula. While the first numerically controlled tools can be dated back to early 1950s, the first digital fabrication laboratory called Center of Bits and Atoms (CBA) was established at Massachusetts Institute of Technology (MIT) in 2001 (Gershenfeld, 2012). Since the year 2000, digital fabrication tools and methods have begun to appear in architecture and engineering education (Blikstein, 2013). Celani (2012) discusses the role of the new digital fabrication laboratories in architectural education. Along with scientific content, she claims that these labs create an opportunity for practical explorations for students. For the integration of VR, Horne and Thompson (2008) present research on the integration of VR within the built environment curriculum, and aim to investigate the role of VR and 3D modelling on the architectural curriculum. This study reports on the integration process and considers how CAM and VR technologies can combine with the existing range of teaching and learning methods. 3. Updating architectural curricula in a digital age Architectural education should produce well-qualified and well-prepared professionals who are familiar with recent technologies. Employers demand that graduates not only have the re-

quired knowledge, but also the appropriate skills to be effective and productive in the workplace. To adapt to these challenges, universities constantly need to review their academic processes, in particular how best to integrate digital technologies, rather than teaching them as supporting technical skills courses, which restricts their creative use as part of the design process. The design studio is still at the core of curricular structure of the schools of architecture and is considered as the norm for architectural design practice. In some architecture programs drawing and representation courses are separated; while in others these courses are integrated into the design studio (Gökmen et al., 2007). Either way students’ need the ability to use these tools creatively. Our intention at Izmir University of Economics is to incorporate digital technology not merely as a tool but as a way of thinking. The first course to introduce CAD/ CAM to students at Izmir University of Economics is “FFD 104 - Computer Aided Technical Drawing” (Varinlioglu et al., 2016). Similar to the first year core Basic Design Studio, this computational course reflects the pedagogy of teaching architectural elements, human scale, and abstract forms without functional requirements. This interdisciplinary approach introduces a discrete studio model, while requiring independent professional courses for each discipline (Duarte et al., 2011). However, an important disadvantage of this pedagogical model is the separation/dissociation of CAD/CAM courses from the main design studio. In the second year, “FFD 201 Computer Aided Architectural Graphics” covers the basics of 3D modelling, architectural graphics, and professional conventions in a one-semester course (Varinlioglu et al., 2017). As this is the second and last of the required CAD courses, it was necessary to incorporate several topics. The course introduces the essential techniques of architectural graphics in two and three dimensions, and stresses their incorporation and application within the virtual technology. Information constituting the initial design created by way of primitive forms exists at the core of the

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

course. Students are not expected to develop their designs any further than the minimum initial requirements. Instead, they are repeatedly encouraged to re-elaborate the initial design within a variety of media, thus remodelling the same information with different tools. 4. Outer space architecture Explorations in outer space have shown significant progress in technological and architectural terms since the 1960s. The first attempts to enable humans to exist and survive in outer space, corresponding to the first wave of outer space development extending to the late 1990s, aimed to create smallscale habitable spaces in the form of a shuttle or a capsule for the specific missions of exploring other celestial bodies, building telecommunication satellites as well as scientific research and photographic data collection. Upgrading the applicable missions in outer space from small-scale scientific research to commercial public scale, the second wave of outer space development opened a new path for space colonisation attracting not only scientists but also private entrepreneurs and ordinary people. Scientific research during the first wave outer space development brought forward a new spatial term called “cabin ecology”, which was used by scholars in astronautics in the late 1950s to describe the environment inside a space vehicle. “The best way to build space cabins”, as the science historian Peder Anker mentions, “was to make their environment as close as possible to the environment found on the surface of the earth” (Anker, 2005: 240). In this context, research on cabin ecology did not initially aim to solve environmental and architectural problems regarding outer space habitats and to build closed liveable environments in space for astronauts, but to construct self-sufficient artificial ecosystems for the use of military forces or as models to handle the ecological crisis on Earth. The Mars One project, which aims to send the first human colonists to Mars in 2032 and establish a permanent human settlement, is an example

of the gradual transformation of outer space architecture from the orbital cabin ecology to on-land outpost architecture. Although the architectural design will primarily be based on adjusting the interiors in accordance with the physiological and functional requirements of the inhabitants, the challenging environmental conditions on Mars as compared to those of Earth may lead to the emergence of new life forms, growth, behaviour and sociability patterns. On Mars, the outpost will expand as more astronauts arrive, creating more living space and ever changing environments for the permanent settlement (Mars One, 2017). In this paper, we consider outer space architecture not as a “cabin ecology” in which outer space habitats are reduced to an interior design problem, but as an outpost architecture of unconventional living forms. In an extra-terrestrial environment where time flow, solar orientation, bodily stimulation and ergonomics as well as microclimatic context is quite different from those on Earth, new forms of human existence and behaviour patterns are likely to occur in and out of the living unit. Such an approach allowed the students to explore a number of unconventional design solutions. Primarily this entailed the use of biomimicry in exploring how these biomimetic organisms could transform themselves. 5. Biomimetic approaches in outer space architecture The British architect Michael Pawlyn defines biomimicry as “mimicking the functional basis of biological forms, processes and systems to produce sustainable solutions” (Pawlyn, 2011: 2). In outer space architecture, biomimicry has rarely been referred to as a design approach. The major reason is that the mainstream approach in designing outer space habitats has always aimed to provide isolated and self-sufficient enclosures for human survival that will resist the challenging unexpected conditions of a relatively unknown environment. In almost all of the architectural solutions for outer space habitats discussed in the previous section, ranging from “cabin ecology” to “outpost design”, the designers

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

have worked within the limited opportunities provided by engineering technologies. Although the Earth, as the only known natural ground where the human species can survive, provides a large variety of biomimetic resources and references for extraterrestrial colonisation, the functional, formal, physical and chemical features of organic life forms on Earth has often been ignored during the design process of outer space habitats. Among various architects who became aware of the importance of space research to ecological design in the 1950s, Buckminster Fuller was one of the key figures who used biomimicry as a model for understanding life on Earth and adapted it into his domes, some of which were designed for military purposes. Based on his geometrical research of spherical trigonometry, he developed and constructed a dome structure with students at the Institute of Design in Chicago in 1949, which later became as “the standard of living package […] for the use of civilians fleeing cities to ‘decentralized communities’ in the event of a nuclear war” (Anker, 2007: 424). In the following years, Fuller constructed a number of domes, a large scale version of a space shuttle interior adapted to human’s bodily features. A complex system of shades was used to control its internal temperature. The sun-shading system was an attempt by the architect to reflect the same biological processes that the human body relies on to maintain its internal temperature. Even more ambitious, Fuller’s original idea for the geodesic dome was to incorporate pores into the enclosed system, further likening it to the sensitivity of human skin (Massey, 2012). 6. The Mars Workshop In order to understand how second year students of architecture at Izmir University of Economics, who developed basic computer skills in their first year, refer to, interpret and utilize digital design tools in the fabrication and representation of an unconventional design context, we conducted a two-week workshop entitled “Mission Mars 2024: A Biomimetic Structural Organism”, as part of the studio course

ARCH 202 in the spring semester of 2017 (Mission Mars, 2017). Originating from the Mars One project initiated by Bas Lansdorp and Arno Wieldersas, students were asked to design a biomimetic structural organism that will function as a habitat for the first colonisers on Mars. The final design was expected to be a standalone organism which can not only structurally withstand the geographical and climatological conditions of Mars, but also grow, move freely, reproduce and/or exterminate itself. As part of the structural organism, the students were also asked to design habitable spaces that will respond to basic human needs such as eating, sleeping, working and socializing. Even though the primary objective of the workshop was to advance their digital design skills in creating unconventional design formations, we considered that in order to propose new growth and living patterns in an extra-terrestrial context, the students also needed a conceptual departure point, and this led to the introduction of “biomimicry” as a supplementary design approach which would initiate their draft proposals. The students worked in teams of six to seven, each of which was assigned in cross-combinations one of four plants: banyan tree, lithops, romanesco and pine cone; one of four animal features: beetle exoskeleton, dragonfly wing, snail shell and moth eyes; and one of the four specific sites on Mars’ surface: Radau, Conches, Maunder and Mistretta (Figure 1). After introductory sessions on advanced digital design tools and biomimicry in architecture as well as Mars’ geography and climate, each team of students was asked to research and analyse their particular plant or animal feature and to build both a digital and a physical model of their given site. In terms of a biomimetic approach, they were asked to design a unique organism that does not simply resemble the plant or the animal feature but mimics and interprets its structural formation. In the initial stage, students presented their design ideas through sketches, diagrams and concept models before they start interacting with the digital tools. Later, students with basic

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

Figure 1. Chart of workshop teams (Halıcı et al. 2017).

Grasshopper Visual Scripting Environment (VSE) skills are encouraged to search for pre-cooked algorithms/ definitions representing their spatial and functional scenarios of the Mars organism, thus gaining the skills to read and understand the algorithms, to guess the potential outcomes generated by these algorithms, to change

parameters to generate more results and also to compound two algorithms for creating their own models. Referring to Oxman’s ontology, teams used digital tools at three different levels: the CAD models, replacement of the physical models with digital ones; the formation models, models showing various stages of the forms/formation;

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

Table 1. Analysis of the design outcomes (GP, growth pattern, UD, unit design, S, structure). The number of dots in each cell refer to the frequency of the biomimetic approach and the digital tools in the design outcomes.

and the generative models, parametric models that can generate new results based on the rules. Four selected examples will be examined in the next section. At the manufacturing phase, the students considered various potential outcomes of the digital fabrication. The 1/5000 site models were produced with laser cuts and 1/200 unit models with 3D printers. But the digital fabrication is not limited to the production of the models. Discovering the potential of 3D manufacturing, they are encouraged to think about the real one-to-one construction of their organisms on Mars. This allowed them to reconsider their concept sketches and ideas, in particular building their colonisation scenarios on Mars, where they frequently utilised digital fabrication. At the representation phase of their design for unconventional environments, students were encouraged to go further than conventional architectural drawings. Using basic VR tools, they were able to simulate the experiential aspects of their work. The finalized 3D digital models were post-processed in an online 3D display and VR display system, with an internal render/ visualisation engine.

7. Analyses of the workshop outcomes We followed a comparative case study method in which we analysed the digital tools and biomimetic approaches of each design team in relation to parameters of growth patterns, unit design and structure particular for the workshop (Table 1) (HalÄącÄą et al. 2017). The initial results show that the use of digital tools is equally distributed among the parameters at different levels of computability, while biomimicry is referred to mostly for parameters of unit design and structure. We consider that the teams which design the same parameters both by means of digital tools and through a biomimetic approach show consistency in their design process in terms of creating an unconventional solution with reference to a conventional living organism. The results have shown that the use of digital tools varies according to the level of design computability. Based on Oxmanâ&#x20AC;&#x2122;s ontology (2008: 106), on the first level are CAD models, which utilize paper-based design methods. Teams 1, 2, 6, 9, 10, 11, 15 utilized digital tools merely as a modelling tool and, team 7 used a paper-based model only, despite digital architectural de-

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

100

sign methods suggesting a new type of vocabulary. The other teams employed digital architectural design methods which had the potential to become the whole process itself, and also present infinite iterations which can be manipulated many times throughout interactive processes. Teams 4, 5, 8 and 13 used parametric tools to create a model capable of showing several forms of their design. However, three examples that particularly stood out are teams 3, 12 and 14, as they showed consistency in their digital design process in terms of creating an unconventional solution with reference to a conventional living organism. They created generative models to create the unit, and to display the growth pattern of the colonisation in Mars. The findings in Table 1 below explain the selected four projects: Teams 3, 12, and 14 all make significant use of digital tools and a biomimetic approach. Although team 6 used biomimicry efficiently without creating generative models, at conceptual level it is worth a mention. It also provides a comparison with team 14, which used the same biomimetic organism. Next, we aimed to establish the extent to which digital tools and biomimicry were reflected in the student’s work, based on the grades awarded, and the personal observations of jury members. There follows an analysis of four successful projects based on the efficiency of using digital tools as a design aid, an appropriate understanding of biomimicry and the exploration of a colonisation program in outer space. Each team was evaluated on the success of using growth patterns, unit design and structural design. While each team of students analysed and interpreted the structural formation of their plant or animal feature as required, the end products varied in terms of the biomimetic approaches that they followed. Some teams focused on the biological structure without directly copying its form but utilizing it to find a functional solution for the long-term survival of users, while some teams transformed the structural form in a more literal way into a living unit. Team 12: Team 12 analysed the extraordinary biological formation of the romanesco which is composed of thou-

Figure 2(a). Presentation boards of Team 12 and 3.

sands of smaller fractal romanescos. The team adapted this formation into a mobile structural organism which can move, rotate, scroll and change its form and place in accordance with sunlight and moonlight. Their structure is an abstracted large scale version of the romanesco, a spiral central module surrounded by conic sub-modules, functioning as living capsules, which enable the whole structure to move and change its position and overall form. In using the growth pattern of romanes-

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

101

Figure 2(b). Presentation boards of Team 12 and 3.

co, they used the generative design approach of L-systems, the algorithm of the spiral. The core is growing in spirals until it reaches its maximum number of units. The fractals of the romanesco plant generated similar sets whose patterns are composed of smaller-scale copies of themselves, processing self-similarity across scales (Lu et al., 2012), and this is reflected in the VSE environment, carrying the biomimetic parameters at a digital level. Thus, they conveyed efficiently the biomimetic

parameters into the form-finding process, by generating alternatives based on the same rules. Team 3: Apart from these biomimetic approaches, there were other teams which did not only integrate the functional and formal characteristics of their reference plant or animal feature, but further proposed unique structural organisms which can exist, grow, reproduce and survive in alternative ways to those of the typical life forms on Earth. In doing so, they gave particular emphasis on the challenging site conditions of Mars. They analysed the lithops plant, which can grow inwards and camouflage itself after extending to a certain degree in accordance with changing degrees of temperature and water vapour level. Taking these characteristics as their starting point, they proposed a structure that provides an organic link between Earth and Mars. According to their scenario, a modular organism created on Earth is sent to Mars and separates into sub-modules that will settle on different areas of the site Radau. As the site conditions allow, the main module will expand, reproduce and spread into smaller modules. The modules which can adapt to site conditions can survive while the others exterminate themselves by growing inwards in weekly periods. The team analysed carefully several characteristics of the plant such as the growth pattern and unit design. The capsules of the plant opening and leading to the new plants led students to develop an algorithm of the fractal geometry for the growth pattern. Each unit, in the form of a sphere, as inspired by Ozdemir and Halici (2016), spreads within the Mars environment using the cellular automata, a generative design approach to define a self-organizing system for managing complex structures within its neighbourhoods (Dinรงer at. Al. 2014). Once one survives, more colonies may spread next to it. If one dies, there is no colonisation around this unit. During construction the fractal joints would help to ease the manufacturing process, where robots will make the camouflage covering under the regolith at the surface of Mars. VR, as a representation tool was expected from each team. For Team 3,

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

102

at the representation level, the conventional physical static models were not enough to reflect their ever changing and growing mechanism. They therefore made a video of the growth pattern and reflected to video mapping to the Mars environment. Teams 6 and 14: Teams 6 and 14 analysed the dragonfly wing as their reference animal feature. The Dragonfly is a flying insect with a centralized body and four equal wings, which can absorb energy from another dragonfly by hitting them. As a unique characteristic, the veins on the wings of the dragonfly are arranged in accordance with the Fibonacci sequence with voronoi pattern, creating a slightly hexagonal pattern growing towards the edges of the wing (Rokicki and Gawell, 2016). Team 6 did not directly copy the form and the pattern of the wing but designed vertical structures which are enveloped by spiral tubes functioning as wind turbines which resist and utilize the strong winds on Mars’ surface. Inside each spiral tube, there are spherical cells continuously moving by wind force. Analogous to the energy absorption characteristic of the dragonfly wing, each cell charges itself as it moves forward and touches another cell. Furthermore, the vertical structures are connected to each other with horizontal tubes, functioning as greenhouses, and arranged in a growing pattern based on the Fibonacci series. Team 14 also analysed the same growth pattern but only in formal terms. They designed a single bowl-like structure that resembles the pattern of the wing and fits onto the volcanic crater in their site Mistretta. Although the two teams differ in their use of digital tools, have both generated promising generative models. Although team 14 achieved a generative model, the algorithmic thinking could not be reflected in the final digital model. Team 6 planned a performance model, where the spiral form would act like a wing of the dragonfly in a vertical direction, but ended up with a simple spiral CAD model. At the growth pattern level, the spiral units would produce energy, using the wind. Although Team 14 took a very

Figure 3(a). Presentation boards of Team 6 and 14.

literal view of biomimicry, they created a growth pattern from the voronoi and ended up with a generative model. The Voronoi pattern divides the space into a number of regions according to the shortest distance to corresponding points in the neighbouring cells (Reinhardt, 2015). Using Delaunay triangulation, the straight line in between the voronoi cells represents the centers of the living units. Similar to the application of voronoi diagrams in contemporary architecture and town planning

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

103

Figure 3(b). Presentation boards of Team 6 and 14.

presented by Nowak (2015), the vessels of the dragonfly would turn into the circulation axes in their colony. The proposals were well adapted to the geography, growth patterns and topographic parameters, and they were successful in terms of using digital tools. 8. Concluding Remarks The outcomes of the workshop have not only shown that the students follow various approaches in designing outer space habitats, ranging from bio-

mimicry to digital tools, but also that they tend to reflect certain preconceptions in their cognitive design process both in formal and functional terms. First of all, even though their design approaches may be unconventional, the actual living units as part of their structural organism mostly take elliptical or spherical forms fully or partially resembling a typical space shuttle. This may be due to the fact that they utilised digital tools to develop growth patterns rather than generating the unit designs in this way. Second, none of the teams developed a proposal for the human life cycle inside, outside or in-between these units, considering this functional issue an interior architecture or engineering problem. The ways in which basic human needs and habitation can differ on Mars from those on Earth and how it reflects on the design of habitable spaces are the two key questions which remain unanswered. The high level of enthusiasm displayed for the projects by the students is worth mentioning as it exceeded our expectations. The presentations were expanded to form part of a public exhibition in collaboration with the Space Camp Turkey in Izmir, and the students were greatly motivated by the positive responses they have received. Most of the architecture schools in Turkey teach digital representation tools in the classroom due to the insufficiency of the technical infrastructure and/or lack of experience of the design studio tutors. In this study, we observed that a more integrated approach had the potential to provide novice architecture students with the adaptability which is necessary to understand digital fabrication, virtual reality and digital architectural design processes. Therefore, this study may be considered an attempt to develop integrated pedagogical models for future architecture curricula. Acknowledgements We would like to express our sincere gratitude to Prof. Dr. Serdar Bayari, Süheyla Müge Halıcı, Gözde Damla Turhan, Mehmet Sadık Aksu, as well as part-time lecturers Filiz Keyder Özkan, Özgür Genca, Lale Başarır, Ayşe Bozkurt Karal for their support and

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

104

guidance, and of course the students of Arch 202 during this workshop. References Anker, P. (2005). The ecological colonization of space. Environmental History. 10, 239-268. Anker, P. (2007). Buckminster Fuller as captain of Spaceship Earth. Minerva. 45, 417-434. Blikstein, P. (2013). Digital fabrication and ’making’ in education: The democratization of invention. Edited by J. Walter-Herrmann and Büching. FabLabs: Of Machines, Makers and Inventors. Bielefeld: Transcript Publishers. Celani, G. (2012). Digital fabrication laboratories: Pedagogy and impacts on architectural education. Nexus Network Journal, 14(3), 469-482. Corser, R. (2010). Fabricating architecture: selected readings in digital design and manufacturing. Princeton Architectural Press. Dinçer, E., A., Tong, H., Çağdaş, G. (2014). A computational model for mass customized housing design by using cellular automata. ITU A/Z, 11(2), 351-368. Duarte, J., Celani, G., Pupo, R. (2011). Inserting computational technologies in architectural curricula. Edited by Ning Gu and Xyangyu Wang. Computational Design Methods and Technologies: Applications in CAD, CAM and CAE Education. IGI Global, 390-411. Gershenfeld, N. (2012). How to make almost anything: The digital fabrication revolution. Foreign Affairs, 91(6), 43. Gökmen, H., Sayar, Y., Süer, D. (2007). Yeniden yapılandırma sürecinde mimarlık eğitimine eleştirel bir bakış, Mimarlık, 337, 63-67. Gül, L. F., Simisic, L., (2014). Integration of digital Fabrication in architectural curricula. Paper presented at the meeting of FabLearn Europe 2014, Digital Fabrication Conference, Aarhus University, Denmark. Halıcı, S. M.,Turhan, G. D., Aksu, S. M., Varinlioğlu, G. (2017). Uzay mimarlığında sayısal tasarım ve üretim araçlarının değerlendirilmesi üzerine Mars özelinde bir çalışma. Paper presented at the meeting of XI. Mimarlıkta Sayısal Tasarım Ulusal Sempozyumu

Konferansı MSTAS, Ankara, Turkey 22-31. Horne, M. and Thompson, E. M. (2008). The role of virtual reality in built environment education, Journal for Education in the Built Environment, 3(1), 5-24. Kotnik, T. (2010). Digital architectural design as exploration of computable functions, International Journal of Architectural Computing, 8(1), 1-16. Lu, X., Clements-Croome, D., Viljanen, M. (2012). Fractal geometry an Architecture Design: Case Study Review. Chaotic Modeling and Simulation (CMSIM), 2, 311-322. Mars One Project (http://www.marsone.com/mission/simulation-outpost) Mission Mars Workshop (https:// missionmars2024.wordpress.com/ blog) Massey, J. (2015). Buckminster Fuller’s reflexive modernism. Design and Culture, 4, 325-344. Nowak, A. (2015). Application of Voronoi diagrams in contemporary architecture and town planning. Challenges of Modern Technology, 6(2), 3034. Oxman, R. (2006). Digital design thinking: in the new design is the new pedagogy. Paper presented at the meeting of 11th International Conference on Computer Aided Architectural Design Research in Asia CAADRIA 2006, Kumamoto, Japan, 37-46. Oxman, R. (2008). Digital architecture as a challenge for design pedagogy: theory, knowledge, models and medium, Design Studies, 29(2), 99-120. Oxman, R. Gu, N. (2015). Theories and models of parametric design thinking. Paper presented at the meeting of 33th eCAADe Conference Proceeding. Vienna University of Technology, Austria. Ozdemir, K., Halici, S. M. (2016). Roll SEED Roll: An Architectural Assessment of a Spherical Mobile Habitat for Mars (SEED_Spherical Environment Exploration Device). Paper presented at the 46th International Conference on Environmental Systems. Vienna, Austria. Pawlyn, M. (2011). Biomimicry in architecture. London, RIBA Publishing. Reinhardt, D. (2015). Coral-Colony - from singularities of mathematical

ITU A|Z • Vol 15 No 1 • March 2018 • G. Varinlioğlu, B. Pasin, H. D. Clarke

105

code to relational networks, Architectural Theory Review, 20:3, 350-364. Rokicki, W., Gawell, E. (2016). Voronoi diagrams - rod structure research models in architectural and structural optimization. Mazowsze Regional Studies. 19, 155-164. Varinlioğlu, G., Alaçam, S., Başarır, L., Genca, Ö, Üçok, I. (2016). Bilgisayar destekli teknik çizimde yeni yaklaşım-

lar: Temsil araçları arası dönüşüm. Yapı, 419, 136-164. Varinlioglu, G., Basarir, L., Genca, O., Vaizoglu, Z. (2017). Challenges in raising digital awareness in architectural curriculum. Edited by G. Cagdas, M. Ozkar, L. Gul and E. Gurer. Computer-Aided Architectural Design, Communications in Computer and Information Science, 136-150.

Unconventional formulations in architectural curricula: An atelier on design for outer space architecture

ITU A|Z • Vol 15 No 1 • March 2018 • 107-122

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings Orkan Zeynel GÜZELCİ1, Sinan Mert ŞENER2 1 orkanguzelci@gmail.com • Department of Interior Architecture and Environmental Design, Faculty of Architecture, Istanbul Kültür University, Istanbul, Turkey 2 mert@itu.edu.tr • Department of Architecture, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.60362

Received: October 2017 • Final Acceptance: November 2017

Abstract Various types of information embedded in the built environment or buildings can be measured by using methods such as entropy to give objective, precise and quantitative results. Jury evaluation is a process where buildings are evaluated subjectively without predefined selection criteria, and that criteria are weighted. The model developed in this study investigates the relationship between entropy values calculated for buildings, and the success obtained as a result of the jury evaluation. Since both design and jury evaluation are not dependent on a single factor, the relationship between single entropy values and the success of the projects cannot be questioned. Therefore, the model being developed in this study handles 5 different entropy values calculated according to 5 factors, weighted independently, and finds total entropy values. To achieve similar results to jury evaluation, a non-dominated sorting algorithm for weighting factors was utilized in relation to an inverted U graph. By finding the weighting between the entropy values, the study aims to resolve a parametric foundation for jury evaluation. Within the scope of this study, 24 municipality building projects designed for architectural project competition between 2015 and 2016 in Turkey, and which have received awards have been evaluated. Keywords Entropy, Information theory, Non-dominated sorting algorithm, Municipality building, Architectural competition.

108

1. Introduction Cities, buildings, works of art, or other man-made artifacts have a high-level organized complexity and each can be handled as a source of information. People, in turn, process the transferred information and make it understandable. Glanzer (1958) defines the person (organism) as an “information processing system”. Various subjective and objective methods are used for the assessment of built environment and buildings. While subjective methods evaluate buildings from the aesthetic perspective, based on personal likes and preferences, objective methods focus on the features of the building that can be calculated. The level of complexity of information obtained from sources or the level of uncertainty contained in the information positively or negatively affects the assessment of objects or buildings. As Vitz (1964) stated, the organism (black box) has perceptual or cognitive response tendencies. While a high level of diversity of information coming from the information source leads to a difficulty of understanding, a low level of diversity causes monotonousness and accompanying vapidity. Instead, it is argued that an average degree of irregularity creates positive feedback from people, as well as pleasure (Vitz, 1964; Berlyne, 1974; Saklofske, 1975). On the other hand, Maddox (1990) emphasizes that there is no satisfactory measure of complexity that distinguishes between what he defines as ordered and disordered complexity. Based on various approaches in the literature, it can be argued that it is an open-ended question as to what degree of complexity could create a more comprehensible or positive impact. It is possible to calculate, and make visible, various types of information embedded in built environments or buildings, using methods which give objective, precise and solid results, such as entropy. Entropy may also be used as a measurement method to meaningfully compare different abstract or concrete architectural compositions. The theory of information and relevant discussions address the measurability of aspects such as complexity

Figure 1. 3 Possible functions between subjective response and entropy (Stamps, 2002).

and uncertainty. Entropy, which is an objective method developed to measure complexity and uncertainty, also offers significant potential in enhancing the comprehensibility of subjective tendencies that involve uncertainty. Based on this point, a model has been proposed which aims to ensure the visibility of the relationship between the multiple entropy values obtained by the measurement of different factors of buildings, and success obtained as a result of the subjective evaluation of an architectural competition jury. The assumption is that a jury evaluates projects according to the inverted U graph based on their entropy values, and this constitutes the basis for developing a model and the calculation of weights. The investigation of this relationship seeks answers for the following questions: • What is the impact of entropy on the jury assessments? • Can the result of jury assessments be estimated/predicted according to the building’s entropy value? • Can subjective means of evaluation, like jury assessment, be associated with objective computational models? 2. Entropy Although entropy has first appeared for measuring physical disorder of substance, in 1949 for measuring the disorder in information it was rediscovered by Shannon and Weaver (1949). For instance, in thermodynamics, entropy value of a substance in crystal form

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

109

Table 1. Entropy calculations as based on a simple letter string (Stamps, 2004).

Figure 2. Basic entropy equality (Stamps, 2004).

is less than the value of substance in melted form (Crompton, 2012). In the Shannon’s information theory, quantity of information that is carried with a message is dependent on the number of probabilities of outcome that can be created by that message. In the case of only one probable output, it is not possible to obtain new information from the message. Suppose if the probability of occurrence of X event is less than that of Y event, since occurrence of X event would cause bigger surprise, X event carries more information (Kan and Gero, 2005). 2.1. Main concepts and basic entropy equality To make an entropy calculation, there should be present a factor and level of occurrence of that given factor. For example, the string of “AAAAAAA” is composed of 7 units. In the string, the only factor being observed is “Letter”; the only level of this factor is the letter “A”. Since there are 7 of instances “A” in the string, the frequency of letter “A” is 7/7. When all units composing the string are identical, there is monotony and entropy value of string is zero. On the other hand, in cases like “ABCDEFG” where all units are different, there is a complete variety and entropy has the highest value (Table 1) (Stamps, 2004). The basic equation that is used for calculating entropy value of information is shown in the Figure 2. In the equation, “H” defines the entropy value

that is calculated and contained in each factor and it is determined in terms of “bits”; furthermore, “p” defines the probability of occurrence of a factor (Stamps, 2004; Crompton, 2012). To structure the entropy formula with letters, for ten pieces of elements made of letters (m=10) and for each unit to have 4 values (n=4), 10 letters are randomly created from one to four (A,B,C,D). When it is assumed that resulting product is “ABBCCCDDDD” in the letter string there are present 1 count of A, 2 counts of B, 3 counts of C, and 4 counts of D. In this case it is observed that frequency rate of letters in the string are 1/10, 2/10, 3/10 and 4/10, respectively. Once the probability of each letter is inserted into the equation and summed, entropy value of letter string is calculated as 1,84 bits (Stamps, 2003). Letter A; Letter B; Letter C; Letter D; 0,33+0,46+0,52+0,53=1,84 bits According to the theory of information, entropy assumes its largest value if the pieces making up the whole have equal occurrence. Therefore, irregular series carry more information than the repeating sequential series of symbols (Crompton, 2012). For example; the sequence of “baa baa baa” is a series that is easier to define, comprehend and remember than the sequence “aba aab baa” (Stamps, 2004). Miller (1956), who is a pioneer researcher, interlinked

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

110

Table 2. Total entropy calculation for two separate factors (Stamps, 2004).

entropy and cognition, suggests that the short-term cognition of human beings is limited to 3 bits. Evidently, low entropy eases perception and cognition. Adhering to the findings of Miller (1956), Stamps (2014) has ascertained the characteristics of his designs, with respect to the 3 bits as the upper limit for each factor in his study. A difficulty encountered in measuring entropy, is the need to calculate according to multiple factors that are completely independent from each other. In cases where factors are independent from each other, total entropy is equal to the sum of entropy values measured for factors. In Table 2, the entropy values of strings consisting of letters are calculated according to two factors that are independent from each other. In the two examples provided in Table 2, the sequences have two factors consisting of letter and font features. In the strings where all units are written in “Times New Roman” font and the letter is “A”, the entropy value is zero since there is no variance. On the other hand, in cases where each unit is featured with a different letter and font, entropy value is obtained for 2.8 bits letters and 2.8 bits fonts. Thus, total entropy value is calculated as 5.60 bits (Stamps, 2004). 2.2. Literature on entropy and architecture The concept of entropy has been used for decades in various fields such as architecture and planning. For instance, it has been used in the calculation of measurable physical features of art works (Arnheim, 1971), building facades (Krampen, 1979), building silhouettes (Stamps, 1998; Stamps, 2004), site plans (Stamps, 2004), abstract compositions produced by LEGO (Stamps, 2012), important buildings reproduced by LEGO (Crompton, 2012) and urban silhouettes (Bostancı, 2008). The built environments, buildings or abstract

objects examined in the studies, were addressed as sources of information and entropy values were calculated according to different factors. Stamps have investigated the relationship between the level of entropy and visual diversity originating from the physical characteristics of the building and the level of satisfaction and dissatisfaction. Various correlations which numerically differ were found between the values calculated according to factors such as form, color, silhouette and facade elements, and the level of pleasure (Stamps, 2002; Stamps, 2003). Objective methods have been developed for measuring complexities and similarities of architectural drawings. Corner points of external contours of drawings are represented with different letters according to vertex characters and entropy values are measures as per the letter string being obtained (Gero and Kazakov, 2001; Jupp and Gero, 2006). It is difficult to determine which sort of a character a stimuli or structure must have to reach a certain level of entropy. Stamps (2014) has developed techniques to produce designs of expected visual complexity and entropy values by listing the parts of the buildings to be used as well as their characteristics. By creating the LEGO models of buildings constructed in different periods, Crompton (2012) attempts to measure the quantity of embedded information according to their shape entropies and benchmark buildings on the basis of the parts which make up the buildings. Stamps (2012), on the other hand, explores the correlation of perceived diversity depending on color and shape with the calculated complexity, using the shape and color features of LEGOs in abstract compositions. Güzelci (2017) has calculated the entropy values of municipality buildings,

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

111

Figure 3. Entropy equation defined in Grasshopper environment.

a customized architectural typology, on the basis of physical characteristics such as solid-void, form and function, investigated whether there is a specific optimum entropy value for different municipality buildings. The study has also investigated whether there is a correlation between the success obtained as a result of a subjective evaluation by the jury of a competition for a municipality building and the project’s entropy value calculated on the basis of quantifiable physical features. 3. Method As observed in the literature that forms the basis of studies for calculation method developed in this study; entropy calculations of buildings and abstract compositions were conducted according to different factors (Stamps, 2004; Crompton, 2012; Stamps, 2012). Furthermore, although entropy calculations in existing literature were made considering single or multiple factors, results seem to have been evaluated independently. The fact that complexity is obtained by summing sub-components does not reflect the internal organization among components (Klinger and Salingaros, 2000). Therefore, the “overall entropy” found by calculating and summing the entropy values dependent on factors applies in cases where there is no relationship, ranking of significance or weight among the factors. Evidently, the municipality building projects focused under the study have complicated design problems with many different inputs. For municipality building designs, many factors aside from aesthetic values, such as functionality, circulation and size of spaces are all considered. However, no findings within the literature state that these factors impact the design equally thus contribute to the sum, when cal-

culating the overall entropy value of the building. No valid conclusion could be reached within the study which explored the relationship between entropy values of the individual factors of the project and the success achieved in the competition (Güzelci, 2017). Building on the results obtained from the previous studies, this study presents a model which multiplies each factors’ entropy values, by given weight coefficients and then explores the correlations with the results of jury assessment. 3.1. Calculation of entropy values of single factors The selected factors in this study are listed, as follows: solid-void, outline, shape, functional distribution and spatial flow (circulation). Generally, all of these factors are present as basic design features which can be examined through all buildings. An algorithmic system was developed to calculate the results without user intervention, with aim to reduce the amount of error at a timely manner for the purpose of assessment on the entropy calculations according to 5 different factors. The algorithm prepared in Grasshopper environment, is capable of recognizing all polygons, surfaces and letters drawn on different layers in two-dimensional drawing software programs, and using them as input in entropy calculations. First, the main entropy equation formula was defined as a mathematical transaction in Grasshopper environment to be able to calculate entropy. Entropy calculation is conducted according to the probability of occurrence of event A, which can be defined as quantitative values in Grasshopper interface, in the space of event B (Figure 3).

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

112

Figure 4. Example of solid-void entropy calculation.

3.1.1. Solid-void entropy (Factor 1) Crompton (2012) has formed the Empire State Building on a three-dimensional grid with size of 8x7x50 by using repeating cubes. After creating the structure entropy values are calculated as per the probability of a cell’s being full or empty as being randomly selected among 3600 cells on the three-dimensional grid. To be able to calculate the solid-void entropy, it is necessary to know the ratio of occupied and empty areas in all floor plans to the total building area. In the study, atriums, staircases, lifts and corridors were regarded as void areas, whereas all other spaces separated by door were regarded as solid areas. As seen in the figure; total area of all floors of the building, solid areas and void areas were found in Grasshopper environment (Figure 4a, 4b, 4c). “Solid entropy” value was calculated by inputting the entropy formula the ratio of occupied areas to total building area, and “void entropy” value was then cal-

culated by inputting the ratio of empty areas to the total building area. The “solid-void entropy” value of the building was calculated by summing these two values (Figure 4d, 4e). This sample calculation was conducted such that all floor plans were considered altogether. It is also possible to make this calculation separately for each floor plan. 3.1.2. Outline entropy (Factor 2) Stamps (1999, 2004) used the number of turns in silhouette lines to calculate the complexity created by building silhouettes and the entropy value. The method used for building or urban silhouettes consisting of two-dimensional lines in that study were adapted to the floor plan contours consisting of two-dimensional lines. To be able to calculate outline entropy, it is necessary to know the number of turns / vertices of the internal and external contours of the floor plans of projects. To calculate outline entropy, first, points are added to all vertices of the

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

113

Figure 5. Example calculation of building outline entropy.

outer and inner contours forming the building’s floor plans, and thus vertex numbers are obtained (Figure 5a). Then, the points overlapping as a result of the superposition of the plan contours of all floors are deleted, and the number of vertex points not overlapping are summed (Figure 5b). To not complicate calculations and therefore ignore the points which do not overlap but are unnoticeably close to each other, a tolerance value in centimeters, which can be changed by the help of a “slider component” is determined (Figure 5c). By dividing the number of all vertex points which do not overlap, by the total number of floors, the average turn/ vertices number per floor is calculated. Finally, by calculating the logarithm in base two for average number of turns on the floor plan, “outline entropy” value is found. (Figure 5d). 3.1.3. Shape entropy (Factor 3) Crompton (2012) calculates shape entropy of buildings remodeled with LEGOs. While making the calculation, considerations of the repetition number of LEGO parts within the total LEGO space, he identifies entropy values of parts’ each individual part and

the total building by summing them. In the municipality building projects being investigated within the scope of study, units such as manager rooms, secretary rooms, meeting rooms, service areas, toilets were found to be repeated with unchanged size and form. In a building formed by repeated parts, “shape entropy” can be calculated according to the irregularity created by the forms of spaces in the building. To be able to calculate shape entropy, it is necessary to identify how many different units exist and the quantity of each unit. In complex buildings consisting of many spaces, such as a municipality building, it is difficult to count this manually. Therefore, an algorithm was developed to count and distinguish different shapes from each other. First, points were assigned to the vertices of spaces drawn in the “shape” layer. According to the vertex number of each form, the edge numbers of shapes were found, and shapes were classified according to their edge numbers. Once the shapes were grouped according to their edge numbers, matching could be done more accurately by calculating edge lengths and areas as well (Figure 6).

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

114

Figure 6. Grouping spaces according to their corner numbers.

Figure 7. Example calculation of average shape entropy.

The entropy value of an individual shape is found according to the rate of incidence of a shape selected from the grouped shapes within the total space of shapes used in the project. By multiplying the entropy value found for the selected shape and the number of repetition of that shape, total amount of complexity added by that shape to the building is calculated (Figure 7a, 7b, 7c, 7d). Upon completion of this pro-

cess for all parts, total entropy is divided by the sum of all shapes making up the building to calculate the building’s “average shape entropy” (Figure 7e, 7f). In the sample calculation shown in the Figure 7; 22 units were calculated with size of 5 m2 and equal number and length of edges, in the first lines of panels. Each of the 22 units with 5 m2 is taken in the total space of shapes seen in the project, it has a value of

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

115

Figure 8. Example of functional distribution entropy calculation.

3.32 bits. Since there are 22 units with a value of 3.32 bits, the multiplication of two values bring a shape entropy of 73.08 bits to the system. In the Figure 7, the sum of entropy values in the list given on the far right-hand-side was calculated as 557.35 bits. By dividing the total entropy value of 557.35 by the total number of shapes, which is 194, the building’s average shape entropy value was calculated as 2.87 bits. The approach followed in this study differs from both the study of Stamps (2012), who made a design using the LEGO vocabulary he developed himself, and from that of Crompton (2012) who determined the entropy values of pieces by searching the whole LEGO space. Since each project investigated in this study was assessed independently, the vocabulary of shapes and distribution of shapes are subject to variance. Based on that distribution the entropy values of parts and the overall systems also vary.

3.1.4. Functional distribution entropy (Factor 4) Stamps (2003) calculated color entropy according to the distribution of colors in two-dimensional abstract compositions, designed with forms of different colors. In this study, using a similar approach, “functional distribution entropy” was calculated on a color canvas created by representing functions of municipality buildings with different colors on the plan. The minimum and maximum entropy values that can be calculated change as the number of functions increases or decreases. In this case, to be able to calculate the functional distribution entropy value and compare the functional distribution entropies of buildings using a standard method, buildings must have equal number of functions and same functions. After reviewing different municipality building projects, 5 main functions encountered in

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

116

Figure 9. Example of spatial flow (circulation) entropy calculation.

all buildings were identified. As part of the study, these five functions were listed as offices, meeting rooms, archives, service areas and circulation areas. After functions are represented by different colors on a two-dimensional drawing environment, the areas of spaces where each function has been seen were summed independently. The entropy value for each function was calculated on the basis of the ratio of the sum of the area of all units in a given function to the total area of the building. By summing the entropy values calculated separately for each of the five functions, the building’s functional distribution entropy was found (Figure 8). 3.1.5. Spatial flow (Circulation) entropy (Factor 5) Stamps (2004) using letters encoded as the five spatial flow elements he has developed on the site plan, Stamps

then represented the spatial flow as a sequence of letters. Apart from the closed spaces for which shape entropy was calculated, the municipality buildings reviewed contained spatial flow elements such as halls, corridors, service corridors, elevators, fire stairs, stairs, waiting areas, inner corridors, stairway landings, foyers, bays, and elevator entrances, which related to each other physically and visually. These units could vary in shape and size. Therefore, spatial flow elements were represented by letters instead of shapes, and the flows in floors were represented by sequences of letters. Thus, it is possible to calculate the circulation complexity embedded in a floor or in the whole building, on the basis of a sequence of letters. By summing the spatial flow entropies of floors, the building’s overall spatial flow entropy value is calculated (Figure 9).

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

117

Figure 10. Placing projects on an inverted U graph.

3.2. Calculation of entropy values of multiple factors Considering the multiple factors that affect entropy values, it is crucial to understand whether there might be certain relationships between those factors. By means of the developed algorithm, as per 5 different physical features of each project, 5 entropy values were found. With the idea that all factors do not influence jury evaluation equally, it is not possible to simply combine values to obtain the overall entropy of the building. Instead, in the study it is assumed that a competition jury makes subjective evaluations according to the “overall entropy” values calculated based on weights between the factors. Berlyne (1960), Kaplan and Kaplan (1989) and Nasar (1987) emphasize that an average amount of complexity is associated positively with the preferences of people. While monotonousness implies both a low level of diversity and vapidity, chaos refers to high level of diversity and vapidity (Stamps, 2003). The design principles used to create visual diversity are used to avoid monotonousness and chaos, or avoid both of them at the same time (Stamps, 2003). It is thought that an average level of complexity in buildings, building blocks, settlements and cities creates the feeling of unity within variety (Elsheshtawy, 1997; Gunawardena et al. 2015). An increase in positive responses as an entropy value approaches a certain value, and a decrease in pleasure as it deviates from the value in a negative or

positive direction causes a Quadratic (Inverted U) graph. It has been emphasized previously that using irregularity at the correct rate, instead of irregularity levels that may lead to monotonousness or chaos, is critical in ensuring a design that makes a positive impact. Based on the theory that an average level of complexity is preferred, it was assumed that the total entropy value of projects selected by the jury as winner always have a level of complexity closest to the average. In this case, the entropy value of the winning project must be closest to the peak (which specified as 12.5 bits) of an inverted U graphic. The second and third place projects are intended to be on the positive or negative side of the x axis, on the condition that they are close to the peak. The honorable mention prizes would be located further from the peak, relative to the second and third projects (Figure 10). As seen in Table 3, by multiplying each factor’s entropy values of all municipality building projects consistently regardless of different competition settings by constant coefficients (w1, w2, w3, w4, w5), and then summing them, a building’s overall entropy value is found. A non-dominated sorting algorithm which has been developed, multiplies 5 entropy values calculated for each project by 5 constant weighting values. While the project’s entropy values calculated according to these factors (a, b, c, d, e) remain constant, weight coefficients (w1, w2, w3, w4, w5) are constantly changed by the genetic algo-

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

118

Table 3. Calculation formula of overall entropy values.

Figure 11. Layers representing the factors.

rithm to conduct the multiplications. By making changes in the weighting values, the algorithm seeks to bring the overall entropy value of a project to the intended point (on the inverted U graph) of evaluation made by the jury. 4. Case study Within the scope of this study, 24 municipality building projects have been selected, which have been awarded at 4 national architectural project competitions organized in Turkey. In the scope of this research, the architectural projects of the project competitions below have been investigated: • Konak Municipality Building Architectural Project Competition (2015) • Efeler Municipality Building Architectural Project Competition (2016) • İnegöl Municipality Building Architectural Project Competition (2016) • Van İpekyolu Municipality Building Architectural Project Competition (2016) The documents relating to the competition projects were obtained

through individual interviews with the project owners. Reasons for choosing municipality buildings were: • Each building has similar organization schemes and functions, • Buildings are composed of repeating parts (units) with the aim of meeting functional requirements also addressing the project brief, • Easy comparison of architectural approaches, schemes, functional distributions and similar features (due to the fact that competition projects are designed according to the same architectural typology and similar briefs). In the first phase of the case study, the measurement of all projects was done according to the 5 physical factors explained in previous section. With the help of the developed algorithm, the drawings which represent different building features (factors) with layers were interpreted in Grasshopper environment and 5 entropy values for each project were calculated (Figure 11). The respective projects investigated within the scope of the research, there was no quantitative similarity observed between measured entropy values such

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

119

Table 4. Calculated and remapped entropy values of all projects.

as solid-void, outline, shape, functional distribution and spatial flow (circulation) entropy. For instance, the solid-void entropy value can be between 0 and 1 bits as well as functional distribution entropy varies between 0 and 2.32 bits. On the other hand, outline, shape and spatial flow entropy values can have a wide range of values, depending on complexity in the building contour, numeric distribution of units, and distribution of spatial flow elements. Due to significant differences between numeric entropy values and weightings of factors on the subjective evaluation process, to reach an overall entropy value calculated with the consideration of all factors; therefore, the simple addition of entropy values would not produce accurate and meaningful results. To solve this problem, the minimum and maximum entropy values calculated according to the factors were remapped between 0.1 and 1. For instance, the maximum shape entropy value 5.54 corresponds to 1, while the minimum value of 3.58 bits corresponds to 0.1. This remapping operation was repeated for all factors. Thus, it was ensured that the multiplied and summed values became comparable (Table 4).

Since optimization was done collectively for the 5 factors for each of the 24 projects, the objective of the genetic algorithm was defined as follows: Multiply the entropy values of each project by the same coefficients/weighting, and bring them to the intended position (points) on the inverted U graph, according to their overall entropy values. For the inverted U graph, a range of entropy values between 0 to 25 bits were set. In this case, the genetic algorithm changed the weighting values to approximate the most successful projects to 12.5 bits. The weighting values fixed by the genetic algorithm after countless generations and multiplication are as follows: 3.33, 3.84, 3.19, 2.27, 2,92 (Figure 12; Table 4). As shown in Table 4, and as stated in the aims, the genetic algorithm failed to bring all the first-prize-winning projects close to the 12.5 bits overall entropy value. The algorithm searches weighting values and overall entropy values, not only for the first projects, but also for the projectâ&#x20AC;&#x2122;s overall. This step aims to place all the projects in the desired order as close as possible. Therefore, calculated results are optimal values for all projects and factors, rather than focusing on the perfection of a given project. Weighted overall en-

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

120

Figure 12. Searching for weights of factors with galapagos.

Figure 13. Outputs of optimization values represented on inverted U graphs.

tropy values of 4 different competitions are illustrated on the inverted U graphs in Figure 13. 5. Conclusion We presented a methodology and the results of an analysis of a group of architectural projects that had previously been evaluated by juries in national competitions. Jury evaluations include many criteria, some of which may not be calculated. In this study, we solely focused on factors that can be digi-

tized. Our analysis was then concerned with how the selection criteria may be weighted. By calculating the weightings, it became possible to determine which factors entropy was more dominant in the evaluation process. When competitions and projects are handled independently, weighting of criteria are changed from one competition to another, relating to jury evaluations. However, we suggest that examining a large number of award-winning projects in different competitions within

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

121

the same framework may be useful in obtaining important findings about selection criteria. Non-awarded projects were not included in the calculation and this can be justified by the following: the tendency of the genetic algorithm reduces the weighting values significantly in order to place the non-awarded projects at the bottom in the ranking scale. In this case, the algorithm can not function properly because there are two conflicting objectives. Developed automated algorithm proves that entropy values calculated based on various single factors differ from one project to another. In addition, the overall entropy values obtained by directly summing the calculated entropy values are not similar. Based on the idea that the average entropy values will have a positive effect, we aimed to rank the entropy values of the projects by multiplying with coefficients with the use of a non-dominated sorting algorithm. The projects ranked according to the 5 coefficients found by the genetic algorithm show similarities in the ranking of the jury evaluation. For this reason, instead of perfectly ranking the projects awarded in a single competition, we have tried to rank all the projects participating in the 4 competitions, concurrently. Figure 13 illustrated that the projects which received the honorable mention were far from the peak, and the entropy values of the projects which won the first three awards were closer to 12.5 bits. It is envisaged that the predictability of jury evaluations will increase if the number of factors in this study is increased. To solve the subjective basis of the jury evaluation, it is not possible to present absolute findings in the research according to the 5 criteria. Increasing the number of factors and number of competitions (which were limitations within this study) may increase the accuracy of the analysis and predictability of the competition results. As a result of this study, certain weighting values have been obtained. In the study, it was found which factors are influential to which degree, in the case which a jury makes selection

as per entropy values by using coefficients. In an application based on another sample group (competition), the weighting may differ depending on the entropy values of the projects and the evaluation criteria of the jury. To conclude, this method is able to make a design evaluation using entropy measurements. The entropy calculation method developed in the scope of this study can be implemented for building such as hospitals, cultural centers, and schools which are composed of rational or complex forms. This study differs from the previous studies, as it uses multiple interrelated entropy values, aiming for the evaluation of a specific architectural typology. This method can be used for the optimization and analyses of designed projects. The algorithm can also make predictions about the potential of project winning a prize in a competition. In future work, the entropy values of the various features of the facade and the properties of the three-dimensional spaces can be calculated. Thus, an entropy value can be obtained that takes into account more features of the building. References Arnheim, R. (1971). Art and Entropy: An Essay on Disorder and Order. Berkeley: University of California Press. Berlyne, D. E. (1960). Conflict, arousal, and curiosity. New York: McGraw-Hill. Berlyne, D. E. (1974). The new experimental aesthetics. In D. E. Berlyne (Ed.), Studies in the new experimental aesthetics (pp. 1-25). Washington, DC: Hemisphere. Bostancı, S. H. (2008). Kent Siluetlerinin Entropi Yaklaşımı ile Değerlendirmesi. (Unpublished doctoral dissertation) Istanbul Technical University Graduate School Of Science, Engineering And Technology, Istanbul. Crompton, A. (2012). The entropy of LEGO®. Environment and Planning B: Planning and Design, 39(1), 174-182. Elsheshtawy, Y. (1997). Urban complexity: toward the measurement of the physical complexity of street-scapes. Journal of Architectural and Planning

A design evaluation model for architectural competitions: Measuring entropy of multiple factors in the case of municipality buildings

122

Research, 14, 301-316. Gero, J. S. & Kazakov, V. (2001), Entropic-based similarity and complexity measures of 2D architectural drawings. In J. S. Gero, B. Tversky and T. Purcell (eds.), Visual and Spatial Reasoning in Design II, Key Centre of Design Computing and Cognition, (pp. 147-161). Sydney: University of Sydney. Glanzer, M. (1958). Curiosity, exploratory drive, and stimulus satiation. Psychological Bulletin, 55(5), 302. Gunawardena, G. M. W. L., Kubota, Y., & Fukahori, K. (2015). Visual complexity analysis using taxonomic diagrams of figures and backgrounds in Japanese residential streetscapes. Urban Studies Research, 2015(173862), 1–12. Güzelci, O. Z. (2017). Investigating the role of Entropy in Design Evaluation Process: A Case Study on Municipality Buildings. In G. Çagdaş, M. Özkar, L.F. Gül, E. Gürer (Eds.) Proceeding of 17th International Conference, CAAD Futures 2017: Future Trajectories of Computation in Design (pp.211-224). Turkey: Istanbul Technical University. Jupp, J., & Gero, J. S. (2006). Visual style: Qualitative and context-dependent categorization. AI EDAM, 20(3), 247-266. Kan, J. W., & Gero, J. S. (2005). Can entropy indicate the richness of idea generation in team designing?. In A. Bhatt (Ed.), CAADRIA’05, Vol. 1 (pp.451-457). New Delhi. Kaplan, R., & Kaplan, S. (1989). The experience of nature: A psychological perspective. New York: Cambridge University Press. Klinger, A., & Salingaros, N. A. (2000). A pattern measure. Environment and Planning B: Planning and Design, 27(4), 537-547. Krampen, M. (1979). Meaning in the urban environment, London: Pion Limited. Maddox, J. (1990). Complicated measures of complexity. Nature, 344, 705. Miller, G. A. (1956). The magical number seven, plus or minus two:

some limits on our capacity for processing information. Psychological Review, 63(2), 81-97. Nasar, J. L. (1987). The effect of sign complexity and coherence on the perceived quality of retail scenes. Journal of the American Planning Association, 53(4), 499-509. Saklofske, D. H. (1975). Aesthetic complexity and exploratory behavior. Perceptual and motor skills, 41(2), 363368. Shannon, C. E. & Weaver, W. (1949) The Mathematical Theory of Communication. University of Illinois Press, Urbana. Stamps III, A. E. (1998). Complexity of architectural silhouettes: from vague impressions to definite design features. Perceptual and motor skills, 87(3_suppl), 1407-1417. Stamps III, A. E. (1999). Physical determinants of preferences for residential facades. Environment and Behavior, 31(6), 723-751. Stamps III, A. E. (2002). Entropy, visual diversity, and preference. The Journal of general psychology, 129(3), 300-320. Stamps III, A. E. (2003). Advances in visual diversity and entropy. Environment and Planning B: Planning and Design, 30(3), 449-463. Stamps III, A. E. (2004). Entropy and visual diversity in the environment. Journal of Architectural and Planning Research, 21(3), 239-256. Stamps III, A. E. (2012). Commentary on the entropy of LEGO®. Environment and Planning B: Planning and Design, 39(1), 183-187. Stamps III, A. E. (2014). Designing for Entropy -1 - Creating stimuli with designed amounts of discrete Shannon information entropy. Retrieved January 1, 2017 from https://www.researchgate.net/profile/Arthur_Stamps/ contributions Vitz, P. C. (1964). Preferences for rates of information presented by sequences of tones. Journal of Experimental Psychology, 68(2), 176-183.

ITU A|Z • Vol 15 No 1 • March 2018 • O. Z. Güzelci, S. M. Şener

ITU A|Z • Vol 15 No 1 • March 2018 • 123-134

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings Ceyda SARICA1, Ebru ÇUBUKÇU2 1 ceyda.sarica@karsiyaka.bel.tr • Karşıyaka Municipality, Plan and Project Management, Izmir, Turkey 2 ebru.cubukcu@deu.edu.tr • Department of City and Regional Planning, Faculty of Architecture, Dokuz Eylul University, Izmir, Turkey

doi: 10.5505/itujfa.2018.57855

Received: July 2016 • Final Acceptance: October 2017

Abstract Preference for ‘color combinations’ have received remarkably little empirical attention and no study compared people’s responses to ‘abstract color combinations’ and ‘color combinations in urban settings’. This study aims to fill this gap and focuses on color combinations rather than isolated colors. 22 color compositions (11 abstract graphics + 11 simulated urban settings) were created. Color compositions included analogous and complementary hues, warm and cool hues, low (5 hues) and high (10 or 11 hues) diversity color compositions. 104 participant evaluated color compositions for (1) arousal, (2) naturalness, (3) relaxation and preference for various objects and settings including (4) clothing, (5) bathroom walls, (6) mall indoors, (7) restaurant indoors, (8) house indoors, (9) building exteriors and (10) any type of object, using a 7-point bipolar scale. The results showed that; (1) color compositions of abstract graphics and pictures of simulated urban settings were rated similarly for ratings of naturalness and preference for any type of object and setting, (2) low and high diversity color compositions were rated similarly for all scales except preference for house indoors, (3) analogous and complementary color compositions were rated similarly for all scales except preference for clothing, (4) warm and cool color compositions were rated similarly for all scales except preference for bathroom walls. The applied value of these results and areas for future research are discussed. Keywords Color combination preference, Computer simulations, Environmental aesthetics, Environmental perception, Environmental psychology.

124

1. Introduction Several studies investigated the most liked colors for various objects. However, only a limited number of them focused on urban environments. There are three reasons to study color preference in urban settings. First, scientific knowledge (and also common knowledge) suggest that the most and the least liked colors for one object may not be similar to that of another. Put it differently, the most liked color for a car or a sofa would be different than that for a building façade. Second, urban environments offer a variety of colors. Yet, most research focused on people’s emotions for ‘isolated’ colors. This lack of interest on color combinations highlights the necessity to understand people emotions to color combinations (rather than isolated colors) in urban settings. Third, color preference studies tend to use color samples rather than contextual colors. Studies comparing people’s responses to abstract colors (e.g. color samples) and contextual colors (e.g. pictures of objects and settings) produced inconsistent results on whether abstract colors are good representatives of contextual colors. Such inconsistencies call for more research on the use of abstract and contextual colors in understanding people’s preference for color combinations in urban settings. In brief, colors in urban environments influence people’s judgments of environmental quality. Yet, little is known about how people evaluate color combinations in urban settings. Put it differently, color combinations have received remarkably little empirical attention and no study compared people’s responses to ‘abstract color combinations’ and ‘color combinations in urban settings’. This study aims to fill this gap by (1) investigating people’s preference for various color combinations; including analogous and complementary hues, warm and cool hues, few and more hues, and (2) comparing people’s evaluations of “abstract color compositions” and “contextualized color compositions - pictures of simulated urban settings”. In terms of environmental aesthetics building exterior color is an important attribute that influence environmental

experience and aesthetic evaluations (Nasar, 1988). Although color is an integral part of design process, environmental coloration is usually practiced in an ad-hoc manner without scientific approach (Smith, 2003). Designers tend to rely on natural talent or practical knowledge that comes from ‘learning by doing’ or ‘trial and error’ (Janssens, 1996). Given that, one could not deny: scientific knowledge may prevent unpredicted, unintentional and costly mistakes. Separate from environmental aesthetics literature, a voluminous number of color research have investigated whether people tend to like some colors over others (Kreitler & Kreitler, 1972; Whitfield & Wiltshire, 1990). Studies revealed certain amount of agreement for the most and the least liked colors. People tend to like blue (Whitfield & Wiltshire, 1990; Camgoz et. al., 2002; Crozier, 1999; Eysenck, 1941; Granger, 1955; Guilford & Smith, 1959; Helson & Lansford, 1970; Hogg et. al., 1979; Saito, 1994; Valdez & Mehrabian, 1994) and dislike yellow (Eysenck, 1941; Guilford & Smith, 1959; Helson & Lansford, 1970; Saito, 1994; Valdez & Mehrabian, 1994). However, such empirical work have extensively used colored chips (samples) and ignored the importance of context (Chin, 2012). There have been only few color evaluation studies on building interiors (Acking & Kuller, 1972; Hogg et. al., 1979; Kuller & Mikellides, 1993; Kwallek, 1996; Kwallek et. al., 1996; Kwallek et. al., 2007; Slatter & Whitfield, 1977; Stahre et. al., 2004; Stansfield & Whitfield, 2005; Stone & English, 1998) and even less on building exteriors (Janssens, 1996; Cubukcu & Kahraman, 2008; Janssens, 2001; O’Connor, 2006; O’Connor, 2011; Kuller, 1996; Sivik, 1974). Put it differently, little is known about how people evaluate urban settings. Common knowledge suggests that the most and the least liked colors in general (or for specific objects) may not apply to color preference in urban settings. A number of studies have been devoted to compare evaluations of abstract colors and contextual colors. For example the most popular colors for women’s fashion were compared to that of residential interior (Stansfield &

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

125

Whitfield, 2005). Similarly, color preferences for color chips (or samples) were compared to pictures of automobile colors (Saito, 1983), colored objects (Taft, 1997) (eg. furniture, bicycle, and computer), building interiors (Hogg et. al., 1979; Ural & Yilmazer, 2010), and exteriors (Sivik, 1974). Some of these studies revealed consistencies, while the others revealed contradictions between evaluations of abstract and contextual colors. Such inconsistencies between findings of various studies may stem from methodological differences. Yet, these contradictory findings call for more research on comparisons of abstract and contextual color evaluations, especially in the context of urban environment. Moreover, colors always exist with other colors. Yet, most studies on color emotions have focused on evaluation of a ‘single (isolated) color’. Similarly, research on building exterior colors tended to investigate single color applications on a single building (O’Connor, 2011). In such studies building exterior color was manipulated via a digital imaging software, to control hue, saturation, and brightness of that façade (Cubukcu & Kahraman, 2008) and to control color harmony with its surrounding (O’Connor, 2006). Only a limited number of studies investigated color emotions for ‘color pairs’ (Ou et. al., 2004) and a few were focused on color combinations (Ural & Yilmazer, 2010). For ‘color pairs’, investigators argued that some emotions (e.g. warm versus cool, hard versus soft) for a color pair could be predicted by averaging individual color scores. However, empirical evidence also suggested that such predictions are not applicable to evaluative scales such as preference (like versus dislike). People’s attitude towards ‘color combinations’ are more complex. In general, people tend to prefer harmonious colors. However, the explanations for harmonious colors are confusing. According to some theorists similar hues (analogous hues) would produce harmonious colors, while for others contrasting colors could also produce color harmony as long as they complement each other (O’Connor, 2011). In brief, color emotions for ‘color combinations in urban

context’ have received remarkably little empirical evidence. This article hopes to pave the way for such research. Recently, Ural and Yilmazer (2010) investigated whether people’s perception for color combinations for indoor settings vary when color combinations are presented via different visualization techniques. The visualization techniques included color chips, abstract compositions, perspective drawings, and three dimensional (3D) models. The results showed poor associations between the semantic ratings of ‘color chips’ and other media and significant associations between ‘abstract compositions’, ‘perspective drawings’ and ‘3D models’. Thus, the authors argued that abstract compositions are good representatives of architectural coloring. Inspired from that study, the present study aims to focus on color combinations in urban settings and investigate whether people’s emotional response to color combinations differ when color combinations are presented on abstract graphics and pictures of simulated urban settings. Empirical research showed that responses to color photographs accurately reflect on site responses (Cubukcu, 2003; Stamps, 1990). In parallel, colored pictures have been extensively used as representatives of real objects (Saito, 1983; Taft, 1997) and settings (Janssens, 1996; Cubukcu & Kahraman, 2008; Kuller, 1996; Sivik, 1974; Saito, 1983; Taft, 1997) in studies of color emotions. 2. Method 2.1. Color compositions The design of color compositions of abstract graphics and simulated urban settings required a sequential process. First, 22 hues with 150 hue intervals (360 / 15 = 24 hue 900 and hue 2700 were excluded from the sample) were selected from a HSB model color space. In HSB model, any color is represented by a set of three numbers representing hue, saturation, and brightness. Hue values vary from 00 to 3600, each representing a distinct color. Saturation is measured as a percentage from 0% (white) to 100% (fully saturated color). Brightness is measured as percentage from 0% (black) to 100% (fully bright color). For this study various satura-

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings

126

tion and brightness levels were tested for each hue then it is seen that half saturated and fully bright hues produced perceptible hue differentiation and proper building exterior colors. Thus half saturated (50% saturation) and fully bright (100 % brightness) hues were selected. Second, three types of ‘color combinations’ were determined based on hue similarity and the diversity (number) of hues. For the first type, six color combinations were created each of which included five similar (or analogous) hues (either warm or cool hues). For the second type, three color combinations were created each of which included 10 dissimilar (or complementary) hues (both warm and cool hues). For the third type, two color combinations were created each of which included 11 similar hues (either warm or cool hues). This way about half of the compositions included low diversity (5 hues) and the other half included high diversity (10 or eleven hues). Similarly, about one third of the compositions included warm colors, one third of them included cool colors and the rest included both warm and cool colors. Table 1 shows the hues that were used in each color combination. For the color compositions of abstract graphics, eleven compositions were generated as a 8 by 5 checkerboard pattern (2 cm X 2 cm squares). For each color type a random number was assigned to each pixel. For ‘Type 1’, the numbers ranged from 1 to 5 (Table 2) for 5 hues. For ‘Type 2’, the numbers ranged from 1 to 10 (Table 3) for 10 hues. For ‘Type 3’, the numbers ranged from 1 to 11 (Table 4) for 11 hues. The hues in each color combination were associated with these random numbers to apply colors to checkerboard patterns (Tables 2 to 4). For each color combination, these 8 by 5 checkerboard patterns were repeated three times (mirrored and 1800 rotated forms were used in repeats) to achieve a wider differentiation between short and long sides of a rectangle (Differentiation between short and long sides was necessary to simulate a series of building facades in an urban setting). This way, the checker board included 40 (8 X 3) cells on the long side and 5

Table 1. Three types of ‘color combinations’ (including 11 separate combination) were specified based on hue similarity and the number of hues in the combination.

Table 2. The right side shows the random numbers (from 1 to 5) assigned to each cell on the 8X5 checkboard pattern. The left side shows the hues assigned to each cell for color combination #1 as an example. For the remaining TYPE 1 color combinations similar procedure was followed.

Table 3. The right side shows the random numbers (from 1 to 10) assigned to each cell on the 8X5 checkboard pattern. The left side shows the hues for color combination #7 as an example. For the remaining TYPE 2 color combinations similar procedure was followed.

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

127

Table 4. The right side shows the random numbers (from 1 to 11) assigned to each cell on the 5X8 checkboard pattern. The left side shows the hues for color combination #10 as an example. For the remaining TYPE 3 color combinations similar procedure was followed.

Figure 1. Abstract color compositions.

cells on the short side. Figure 1 shows the eleven ‘abstract color compositions’. For the color compositions of simulated urban settings, three types of building silhouettes were designed considering the proportional relations between hues in each type of abstract graphic color combination. Same proportional relations in simulated urban setting color compositions were

achieved by controlling the total area of building façades. To that end, the height and the width of each building façade was manipulated. For 5 analogous hues (see ‘type 1’ in table 1), each hue was presented on two building facades. Thus the composition involved 10 buildings (Figure 2). For 10 complementary hues (see ‘type 2’ in table 1), each hue was presented on one building façade. Thus the composition involved 10 buildings (Figure 2). For 11 analogous hues (see ‘type 3’ in table 1), each hue was presented on one building. Thus, the composition involved 11 buildings (Figure 2). The location of each building in the composition was specified randomly. Figure 3 shows the eleven ‘simulated urban setting color compositions’. Note, the number of hues and the proportion of each color on each of the eleven color combinations were the same in ‘abstract color compositions’ and ‘simulated urban setting color compositions’. However, the adjacency of each color to each other was not controlled between ‘abstract color compositions’ and ‘simulated urban setting color compositions’. This methodological limitation should be accounted in further studies. In brief, there were 22 color compositions (11 abstract graphics + 11 simulated urban settings) and each participant was asked to rate 8 of them. In order to keep participants interest, participants were not asked to rate all color compositions. The compositions that will be evaluated by each participant was selected by stratified random sample. Six sub-groups were determined, each of which involved 8 color compositions. In each group, four were ‘5 analogous’ hues, two were ‘10 complementary’ hues and two were ‘11 analogous’ hues. Also in each group, half of the color combinations were ‘abstract color compositions’ and the other half were ‘simulated urban setting color compositions’. Finally, cool and warm color combinations were equally balanced in each group. 2.2. Survey Studies showed arousal, naturalness, and relaxation are particularly important when studying color pref-

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings

128

erence (Cubukcu & Kahraman, 2008) in physical environments, because studies on environmental aesthetics showed that environmental preference is affected by such emotions (Nasar, 1988). People prefer environments with moderate levels of arousal, high levels of naturalness, and relaxation (see literature review in Cubukcu and Kahraman, 2008). Thus, this study examined people’s emotions of arousal, naturalness, and relaxation in addition to preference. 7-point bipolar semantic differential scales were used to measure arousal (1 = sleepy, 7 = arousing), naturalness (1 = artificial, 7 = natural), and relaxation (1 = distressing, 7 = relaxing). Preference was evaluated in 7 ways using a 7-point bipolar scale (eg. 1 = dislike, 7 = like). Participants were asked to rate their preference on each color combinations for objects [including (1) clothing, (2) any type of object] and for settings [including (3) bathroom walls, (4) mall, (5) restaurant and (6) house indoors and (7) building exteriors]. The survey included questions about participants’ demographic characteristics (age, gender, color deficiency, and the city they grow-up) and their familiarity with colors via three questions. The first question asked whether they are involved (or not involved) in activities related to color such as painting. The second one asked how they evaluate the diversity of color in their environment (simple / diverse / do not know). The third one asked whether they are conservative or flexible in color preference for various objects. For this last question participants were asked to pick one comment among four; (1) I am conservative, I have favorite colors which I tend to use on various objects; (2) I am a little flexible. Put it differently, although my color preference depends on the object, I have favorite colors which I tend to use more often; (3) I am flexible, my color preference depends on the object, and (4) I do not care about colors. 2.3. Participants 104 students studying in Ege University, Geography Department agreed to participate in the study. However, 14 participants were excluded from

Figure 2. The building areas in simulated urban settings were controlled to achieve the same proportion of each color in each type of abstract color compositions. TYPE 1 (5 analogous hues) #1

Warm Colors 1

Warm Colors 2

Warm Colors 3

Cool Colors 1

Cool Colors 2

Cool Colors 3

TYPE 2 (10 complementary hues) #7

Warm and Cool Colors 1

Warm and Cool Colors 2

Warm and Cool Colors 3

TYPE 3 (11 analogous hues) #10 Warm Colors

#11 Cool Colors

Figure 3. Simulated urban setting color compositions.

the sample, 13 for not using corrective equipment (contact lenses or eyeglasses) for their vision deficiencies and 1 for being colorblind. Thus, the results were analyzed for 90 participants between the ages of 17 and 30 (MEAN: 22.58; SD = 2.30). The study group was about balanced as to gender (41% female, 59% male). All participants were university students, and no participant reported having a diverse cultural background; majority (about 30%) spent

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

129

Table 5. Participants’ evaluations (mean scores for various scales) for abstract graphics and pictures of simulated urban settings.

Table 6. Participants’ evaluations (mean scores for various scales) for color compositions with few (5 hues) and more (10 or eleven hues) hues.

most of their life in the third largest city of Turkey, Izmir. Most participants (about 90%) revealed that they are not involved in activities related to color (such as painting). About half of the participants (48%) rated the diversity of color in their living environments as high. For general color preference, results showed that about 18 % were conservative, about 32% were a little flexible, about 31% were flexible, and about 19% revealed that they do not care about colors. 2.4. Procedure Six groups of people (18, 17, 18, 13, 17, and 21) were seated in a classroom and received a brief written and verbal instruction about the task. First they viewed a Mondrian Painting project-

ed onto a screen (which was about 122 cm × 152 cm) and filled the survey. They rated a painting first, rather than the color compositions, to get familiar with the evaluative questions. Then one randomly selected color composition was displayed and participants were asked to fill the evaluative questions (arousal, naturalness, relaxation and preference for various objects and settings) on the form. The color composition was replaced randomly by another until 8 compositions were assessed for all scales. The survey took about 20 – 25 minutes for each group. The participants were not allowed to ask questions to the investigator or to each other during the evaluation of 8 images. 3. Results Participants’ evaluations were compared between different type of color compositions; abstract graphics versus pictures of simulated urban settings, low versus high diversity (number) of hues, analogous versus complementary hues, warm versus cool hues. Color compositions of abstract graphics and pictures of simulated urban settings were rated similarly for ratings of naturalness and preference for any type of object and setting. Both types of color compositions were rated as below average for all scales (means ranged from 2.69 to 3.49). However, the difference between color compositions of abstract graphics and pictures of simulated urban settings achieved a statistical significance for arousal and relaxation scores. Abstract graphics were found to be more arousing and more relaxing compared to pictures of simulated urban settings (Table 5). Color compositions with few (5 hues) and more (10 or 11 hues) hues were rated similarly for ratings of arousal, naturalness, relaxation, and preference for any type of object and setting except house indoors. Both types of color compositions were rated as below average for all scales (means ranged from 2.64 to 3.45). For house indoors, color compositions with few hues were found to be more preferable than color compositions with more hues and this difference achieved marginal significance (Table 6).

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings

130

Color compositions with analogous and complementary hues were rated similarly for ratings of arousal, naturalness and relaxation and preference for any type of setting and object except clothing. Both types of color compositions were rated as below average for all scales (means ranged from 2.67 to 3.41). For clothing, color compositions with analogous hues were found to be more preferable than complementary hues, and this difference achieved marginal significance (Table 7). Color compositions with warm and cool hues were rated similarly for ratings of arousal, naturalness, relaxation and preference for any type of object and setting except bathroom walls. Both types of color compositions were rated as below average for all scales (means ranged from 2.75 to 3.54). For bathroom walls, color compositions with cool hues were found to be more preferable than that with warm hues, and this difference achieved statistical significance (Table 8). 4. Discussion Voluminous number of studies have focused on color preference. Although color preference could be product specific, colors in urban settings received remarkably little empirical evidence. One reason for this lack of interest could be related to the methodological limitations to represent and control the variety of colors in urban settings. This study investigated people’s evaluation of color combinations which are presented in two different forms; (1) pictures of simulated urban settings, (2) abstract graphics. A previous study on architectural indoors (Ural & Yilmazer, 2010) found that abstract compositions are good representatives of architectural coloring. In another study (Guerin et. al., 1994) six abstract color palettes were developed to represent six pictures of interior environments. Based on the results, authors argued that abstract color palettes are valid testing instruments to study meaning of color in interior environments. The present study supported those findings. People evaluated pictures of simulated urban settings and abstract graphics similarly for scales of arousal, naturalness, relaxation, and preference. Recall, questions

Table 7. Participants’ evaluations (mean scores for various scales) for color compositions with analogous and complementary hues.

Table 8. Participants’ evaluations (mean scores for various scales) for color compositions with warm and cool hues.

on preference were specified for different objects (eg. clothing) and settings (eg. mall or house indoors). Results indicate that, people are able to predict the application of color combinations on various objects and settings no matter how the color combinations are presented; either by abstract graphics or pictures of simulated urban settings. Put it differently, when people are asked to imagine color combinations on various objects they are not influenced by presentation technique. In brief this study provide empirical evidence that, abstract compositions are good representatives of simulated urban settings and could be used to understand people’s preference for combinations of colors in outdoor settings. This finding is particularly important for urban designers and environmental psycholo-

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

131

gist who need to understand people’s response to various color combinations in urban settings. They could save time when they use abstract graphics rather than computer models of real world settings. Note however, this study compared only two conditions; abstract graphics and two dimensional simulated urban settings. Future studies should compare abstract graphics with three dimensional real and virtual environments. This study also investigated the influence of diversity of colors on preference. In this study, color combinations included 5, 10 and 11 hues to represent low and high diversity of colors. Results showed that both were rated similarly (below average) for ratings of arousal, naturalness, relaxation, and preference for any type of objects and settings except house indoors. Only for house indoors people tend to prefer less hue diversity. Note, in this study diversity was achieved by hue differentiation in a color combination. Future studies may use saturation and brightness differentiation to test influence of color diversity on preference. Also, in this study color combinations had 5 to 11 hues. Greater differentiation in color compositions (eg. using 3 to 100 hues) may yield different findings. In brief, whether diversity contributes to, or detracts from, environmental visual quality calls for more research. It is widely believed that, in urban environments buildings have to be in harmony with each other (Ünver & Dokuzer, 2002) Color theorists and practioners showed great interest in laws of color harmony (Sivik & Hard, 1994; O’Connor, 2010). Theory and research showed that analogous and complementary colors could produce color harmony. Although this study did not intend to provide empirical support to laws of color harmony (whether analogous or complementary colors produced harmony), it showed that both analogous and complementary colors were rated similarly for ratings of arousal, naturalness and relaxation and preference for any type of object except clothing. Only for clothing, analogous color compositions were found to be more preferable than complementary ones. However note, this study did not

measure color harmony. Thus, future studies should test how people rate analogous and complementary colors in terms of color harmony and how color harmony influences people’s color preference in urban settings remains to be seen. Moreover, in this study analogous and complementary hues were selected from HSB (hue, saturation, brightness) model color space with 150 hue intervals. There are web based tools (such as the color wheel expert, color wheel pro, color wizard see Chin (2012) for a review) to select matching colors. Similarly, Chin (2012) introduced a color selection system with which one can select proper color for building exteriors using a 3D coloring simulation tool for city scenes. Future studies may consider using such tools to design various color combinations that could be considered to be harmonious and inharmonious. Previous studies on color emotions usually grouped colors as warm and cool colors. Studies showed that although color preference varies in time, people tend to prefer warm colors for residential interiors (Stansfield & Whitfield, 2005). Also warm colors are found to be more arousing and less relaxing than cool colors (Yildirim et. al., 2011). Considering the environmental aesthetic literature which argues that people tend to visit moderately arousing and highly relaxing environments and avoid highly arousing and distressing environments, one expects both warm and cool colors to be preferable in urban settings. The findings of the present study showed that warm and cool color compositions were rated similarly for ratings of arousal, naturalness, relaxation and preference for any type of object and setting except bathroom walls. Only for bathroom walls, cool hues were found to be more preferable. Note, this study focused on cool and warm colors to study color combinations in urban settings. However, existing urban settings involve achromatic colors (eg. white) or more variety which cannot be grouped as warm and cool. A useful extension of this study may focus on color combinations in existing urban environments rather than cool and warm colors. It is necessary to highlight the fact that, this pa-

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings

132

per makes no claim to provide concrete evidence relating to the most preferred hues (warm or cool) in urban settings, as urban environments provide a variety of colors. It intends to generate further discussion and research on preference of color combinations in urban settings. In other words, the research is not definitive or conclusive. It aims to pave the way to study color combinations rather than isolated colors. Hard and Sivik (2001) highlight the fact that “the number of colors is very large and the number of possible colour combinations is almost inﬁnite’’ (p 4). Studying color combinations required a systematic way of selecting color combinations. In this study, the selection was based on hue differentiation. The combinations vary according to which they involve (1) warm, cool or both type of colors, (2) analogous or complementary colors, and (3) diversity of colors (5, 10 and 11 hues). In selecting color combinations future studies may consider using Shigenobu Kobayashi’s (Kobayashi, 1981; Kobayashi, 1987) seminal publications on “color image scale”. With more than hundred basic colors he created more than thousand color combinations. The color combinations were then matched with about two hundred semantic concepts like urbane, traditional, modern, and comfortable. Future studies which would test people’s emotional response to Shigenobu Kobayashi’s color combinations in real and simulated urban settings are on call. Finally the methodological limitations related to the experimental set up and the characteristics of the subject group should be addressed. There are three limitations. First, colors’ proportional relations in abstract graphics and pictures of simulated urban settings were controlled but adjacency of colors in two types of presentation techniques were not controlled. Subsequent studies should control adjacency of colors in abstract graphics and contextual presentations. Second, pictures of simulated environments were used to represent an urban setting. Future studies should examine to what extent the evaluation of colors on pictures on a computer screen is relevant for judgments of real urban settings. Third,

the target population of this study was young students in Western Turkey. Whether the results of the present study will apply to different cultures remains to be seen. More work needs to be done to test the generalization of the results to various demographic groups (children, elderly) as well. 5. Conclusion Voluminous number of studies have focused on color preference in general. Although color preference could be product specific, colors in urban settings received remarkably little empirical evidence. One reason for this lack of interest could be related to the methodological limitations to represent and control the variety of colors in urban settings. This study investigated people’s evaluation of color combinations which are presented in two different forms; (1) pictures of simulated urban settings, (2) abstract graphics. Results showed that, abstract compositions are good representatives of simulated urban settings and could be used to understand people’s preference for combinations of colors in outdoor settings. This finding is particularly important for urban designers and environmental psychologist who need to understand people’s response to various color combinations in urban settings. This study is important as it integrates two literatures, environmental psychology and color research in general, and environmental aesthetics and color preference in particular. The methodology derived from color research literature could inspire new research in environmental aesthetics. More research needs to be done to understand the relation between color and environmental aesthetic evaluations. Acknowledgement This paper is derived from the first author’s master’s thesis. That was undertaken in The Graduate School of Natural and Applied Sciences, Department of City and Regional Planning at the Dokuz Eylul University. The authors would like to thank to Dokuz Eylul Univesity, Graduate School of Natural and Applied Sciences and to the students who volunteered to participate in the study.

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

133

Sarıca, C. (2011). Yapılı ÇevredeRenk Algısı: Ampirik Bir Çalışma [Colour Perception In Built Environments: An Empirical Study] (Unpublished master’s thesis). Dokuz Eylul University, The Graduate School of Natural and Applied Sciences, Izmir. References Acking, B. C., & Kuller, H. (1972). The perception of an interior as a function of its colour. Ergonomics, 15(6), 645-654. Camgöz, N., Yener, C., & Güvenç, D. (2002). Effects of hue, saturation, and brightness on preference. Color Research and Application, 27(3), 199-207. Chin, S. (2012). Colour scheme supporting technique based on hierarchical scene structure for exterior design of urban scenes in 3D. Color Research and Application, 37(2), 134-147. Crozier, W. R. (1999). The meanings of colour: Preferences among hues. Pigment and Resin Technology, 28(1), 6-14. Cubukcu, E. (2003). Photo versus on-site surveying to test artwork preference. In: The 34th Annual Conference of The Environmental Design Research Association (EDRA), People Shaping Places Shaping People, Minnesota, USA. Cubukcu, E., & Kahraman, I. (2008). Hue, saturation, lightness, and building exterior preference: An empirical study in Turkey comparing architects’ and nonarchitects’ evaluative and cognitive judgments. Color Research and Application, 33(5), 395-405. Eysenck, H. J. (1941). A critical and experimental study of colour preferences. The American Journal of Psychology, 54(3), 385-394. Granger, G. W. (1955). An experimental study of colour preferences. The Journal of General Psychology, 52(1), 3-20. Guerin, D. A., Park, Y., & Yang, S. (1994). Development of an instrument to study the meaning of color in interior environments. Journal of Interior Design, 20(2), 31-41. Guilford, J. P., & Smith, P. C. (1959). A system of color-preferences. The American Journal of Psychology, 72(4), 487-502. Hard, A., & Sivik, L. (2001). A theory of colors in combination—a de-

scriptive model related to the NCS color-order system. Color Research & Application, 26(1), 4-28. Helson, H., & Lansford, T. (1970). The role of spectral energy of source and background color in the pleasantness of object colors. Applied Optics, 9(7), 1513-1562. Hogg, J., Goodman, S., Porter, T., Mikellides, B., & Preddy, D. E. (1979). Dimensions and determinants of judgements of colour samples and a simulated interior space by architects and non‐architects. British Journal of Psychology, 70(2), 231-242. Janssens, J. (1996). Individual and contextual factors in urban colour perception. In: Sivik L, editor. Colour and Pyschology. AIC Interim Meeting 96. Gothenburg, Sweden. Janssens, J. (2001). Facade colours, not just a matter of personal taste-a psychological account of preferences for exterior building colours. Nordic Journal of Architectural Research, 14(2), 17-21. Kobayashi, S. (1981). The aim and method of the color image scale. Color Research and Application, 6(2), 93-107. Kobayashi, S. (Ed.). (1987). Book of Colors: Matching Colors, Combining Colors, Color Designing, Color Decorating. Japan: Kodansha International. Kreitler H, Kreitler (1972) S. Psychology of the Arts. Durham, NC, USA: Duke University Press. Kueller, R., & Mikellides, B. (1993). Simulated studies of color, arousal, and comfort. In: Marans RW, Stokols D, editors. Environmental Simulation:Research and Policy Issues. New York: Plenum Publishing Corp, Springer. Küller, R. (1997). Preferences for colours in urban places. AIC Interim Meeting Gothenburg, Sweden. Kwallek, N. (1996). Office wall color: An assessment of spaciousness and preference. Perceptual and Motor Skills, 83(1), 49-50. Kwallek, N., Lewis, C. M., Lin-Hsiao, J. W. D., & Woodson, H. (1996). Effects of nine monochromatic office interior colors on clerical tasks and worker mood. Color Research and Application, 21(6), 448-458. Kwallek, N., Soon, K., & Lewis, C. M. (2007). Work week productivity, visual complexity, and individual envi-

Evaluating color combinations using abstract graphics versus pictures of simulated urban settings

134

ronmental sensitivity in three offices of different color interiors. Color Research and Application, 32(2), 130-143. Nasar, J. L. (1988). Environmental aesthetics: Theory, research, and application. USA: Cambridge University Press. O’connor, Z. (2006). Bridging the gap: Facade colour, aesthetic response and planning policy. Journal of Urban Design, 11(3), 335-345. O’Connor, Z. (2010). Colour harmony revisited. Color Research and Application, 35(4), 267-273. O’Connor, Z. (2011). Façade colour and judgements about building size and congruity. Journal of Urban Design, 16(03), 397-404. Ou, L. C., Luo, M. R., Woodcock, A., & Wright, A. (2004). A study of colour emotion and colour preference. part II: colour emotions for two‐colour combinations. Color Research and Application, 29(4), 292-298. Saito M. (1994). A cross-cultural study on color preference in three Asian cities: Comparison between Tokyo, Taipei and Tianjin. Japanese Psychological Research, 36(4), 219-232. Saito, T. (1983). Latent spaces of color preference with and without a context: Using the shape of an automobile as the context. Color Research and Application, 8(2), 101-113. Sivik, L. (1974). Colour Meaning and Perceptual Color Dimensions: A Study of Exterior Colors. Goteborg Psychological Reports, 4 (14). Sivik, L., & Hård, A. (1994). Some reflections on studying colour combinations. Color Research and Application, 19(4), 286-295. Slatter, P. E., & Whitfield, T. W. (1977). Room function and appropriateness judgments of color. Perceptual and Motor Skills, 45(3 suppl), 10681070. Smith, D. (2003). Environmental colouration and/or the design process. Color Research and Application, 28(5), 360-365.

Stahre, B., Hårleman, M., & Billger, M. (2004). Colour emotions in larger and smaller scale. Proceedings of Interim Meeting of the Association Internationale de la Coleur (AIC)“Color and Paints Porto Alegre, Brazil. Stamps, A. E. (1990). Use of photographs to simulate environments: A meta-analysis. Perceptual and Motor Skills, 71(3), 907-913. Stansfield, J., & Whitfield, T. W. (2005). Can future colour trends be predicted on the basis of past colour trends?: An empirical investigation. Color Research and Application, 30(3), 235-242. Stone, N. J., & English, A. J. (1998). Task type, posters, and workspace color on mood, satisfaction, and performance. Journal of Environmental Psychology, 18(2), 175-185. Taft, C. (1997). Color meaning and context: Comparisons of semantic ratings of colors on samples and objects. Color Research and Application, 22(1), 40-50. Ural, S. E., & Yilmazer, S. (2010). The architectural colour design process: An evaluation of sequential media via semantic ratings. Color Research and Application, 35(5), 343-351. Ünver, R., & Dokuzer Öztürk, L. (2002). An example of facade colour design of mass housing. Color Research and Application, 27(4), 291-299. Valdez, P., & Mehrabian, A. (1994). Effects of color on emotions. Journal of Experimental Psychology: General, 123(4), 394–409. Whitfield, T. W. Allan and Tim J. Wiltshire (1990),“Color Psychology: A Critical Review,”. Genetic, Social, and General Psychology Monographs, 116(4), 387-411. Yildirim, K., Hidayetoglu, M. L., & Capanoglu, A. (2011). Effects of interior colors on mood and preference: comparisons of two living rooms. Perceptual and Motor Skills, 112(2), 509524.

ITU A|Z • Vol 15 No 1 • March 2018 • C. Sarıca, E. Çubukçu

ITU A|Z • Vol 15 No 1 • March 2018 • 135-145

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach Meriç DEMİR KAHRAMAN1, Burak PAK2, Kris SCHEERLINCK3 1 demirme@itu.edu.tr • Department of Urban and Regional Planning, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey 2 burak.pak@kuleuven.be • Department of Architecture, Faculty of Architecture, Katholieke Universiteit Leuven, Brussels, Belgium 3 kris.scheerlinck@kuleuven.be • Department of Architecture, Faculty of Architecture, Katholieke Universiteit Leuven, Brussels, Belgium

doi: 10.5505/itujfa.2018.58569

Received: March 2017 • Final Acceptance: March 2018

Abstract Over the recent decades, both the requirements for and the affordances of public spaces have been an unavoidable and growing discussion in the spatial sciences literature. This growing discussion and research have been articulated through the argument that public spaces have been eroding under the neoliberal conditions and the capitalist mode of production. However, from the insights of social sciences, as the physical setting to be included in socio-political life, public spaces appear exclusionary for some as a timeless fact. Although historical public spaces have been idealized and envied, they appear as ideal places for a privileged spectrum of the societies to learn how to rule and to teach the rest how to obey rather than to allow them to be included in the public sphere. Considering the meaning of to be public, this study claims that this is the paradox of public space, which becomes evident in contemporary rising social struggles for public spaces in the form of occupy movements. In this context, this study aims to anatomise the paradox of public space from also the insights of social sciences in the conditions of representative democracy. As the main contribution of this study, we introduce a re-interpretation of Lefebvre’s multi-triads and operationalize his concept of heterotopia to offer a deeper understanding in revealing the paradoxical production of public spaces. We conclude that the social production of a heterotopia is the manifestational realization of an ideal public space and the dissolution its paradox for only a temporary period of time. Keywords Production of space, Public space, Political representation, Heterotopia, Occupy movements.

136

1. Introduction Resource requirements such as land, buBeginning from Agora of the ancient Greeks, public spaces have been the major components of cities for centuries and have also become a major issue for design professionals and researchers especially over the recent decades. Desired features and design parameters for a better public space have been discussed from the domain of spatial sciences repetitiously. However, since the turn of the millennium, especially following the crisis of 2007-2008, a significant amount of critiques also emerged from a realistic perspective emphasising the false romanticization of historic public space and their contestable existence as a representational tool of power, which this study claims that it is the paradox of public space (Robbins, 1993; Iveson, 2007; Madanipour, 2010; Berman, 2012). In parallel with this tendency, the main motivation of this study is to develop these rising critiques further with knowledge from also the insights of social sciences. Based on this motivation, our main research question is: “How can the paradox of public space and its social production be anatomised and dissolved?” To answer this question, in Section 1, we review the recent critiques and reveal that public spaces have always been a matter of struggles between different publics and have been exclusionary places for a large spectrum of the societies as a timeless fact. Discussing the semantics of being public, we develop further the paradox of public space together with discussions on contemporary occupy movements and the paradox of political representation (Pitkin, 1967; Runciman, 2007). In this context, in order to offer a deeper and practical understanding of the paradox of public space and its social production in a given urban pattern, we introduce a re-interpretation of Lefebvre’s (1991; 2003) multi-triads and operationalize his concept of heterotopia in Section 2. In conclusion, we suggest a number of conditions for the paradox of public space to be occurred and dissolved. 2. The phantom of public space and the paradox of political representation Since the 1970s, increasing numbers of studies dealing with spatial scienc-

es have focused on the issue of spatial requirements and importance of public spaces in everyday urban life. In fact, as it is well known, this period overlaps with the abandonment of Keynesian economics in most countries and the rise of neoliberalism. Indeed, this new socio-economic and political turn reflected itself in urbanization eventually, which has been driven by the spatial organization of the brutal capitalist system, especially by means of privatization of urban space. This conjuncture has forced urban designers and architects to re-evaluate both the requirements for and the affordances of public spaces as the main physical settings of everyday life. In accordance, changing typologies and functions of public spaces in a given urban pattern and their publicness has become an unavoidable and growing discussion in the literature. This conjuncture after the 1970s reflects the desired features of public spaces to be achieved in the proposed definitions such as: • composed of the presence of other people, activities, events, inspiration, and stimulation (Gehl, 1987), • a common ground for people to perform functional and/or ritual activities either as a part of daily routine and/or periodic festivities (Carr et al., 1992), • places that belong to a community, not developers/investors or police and traffic wardens (Tibbalds, 1992). The main reason in the background of this growing research and discussion is the argument that public space has been under threat from the modern urban life that is characterized by the capitalist mode of production, consumerism, privatization, restrictive regulations on public space, controlling and militarizing public space through security measures, police force and/ or high tech security camera systems, social exclusion from public space and related politics (Demir Kahraman & Türkoğlu, 2017). This study claims that theoretical idealization of public spaces and defining them as open and accessible for all does not coincide with the practical realities more than ever today. On the

ITU A|Z • Vol 15 No 1 • March 2018 • M. Demir Kahraman, B. Pak, K. Scheerlinck

137

1 Kurt Iveson (2007) used the term “Phantom” after Walter Lippmann (1925) and Bruce Robbins (1993).

other hand, publicness, although it has been told that it has been lost in neoliberal urbanism, has always been a matter of struggles and so public spaces have always been exclusionary places for some throughout history. The roots of the discussion of publicness arose from insights of social rather than spatial sciences, especially since the 1950s under the post-war conditions after the WWII. In social sciences, the term “public sphere” is used to describe aspects that reach beyond the physical limits of publicness, addressing its inherently political nature. To be public, which is essentially making the public realm, is a matter of to be seen and heard in the material human world: to be included in the political life. In other words, every conversation performed by private individuals, who assemble to form a public body compose a portion of the public sphere, which mediates between society and state (Arendt, 1998; Habermas, 1974). Thereby, as the centres of everyday life and the places in which the public sphere has been produced and re-produced, public spaces appear as purely political. Yet, despite the legalization of free speech, press and assembly for everyone, even liberal public sphere and so the space was open to everyone only in principle (Iveson, 1998); in reality, society has been polarized by class struggles, so the public has been fragmented into a mass of competing interest groups; multi-publics (Fraser, 1990). The relations between multi-publics, which are the struggles between dominant and counter publics, produce and re-produce both the public sphere and so space (Demir Kahraman & Türkoğlu, 2017). Having this concern, Madanipour (2010) notes that the claims of different publics over space contest the others, and the struggle between them causes a simultaneous process of inclusion and exclusion, and this is why he asks “Whose public space?” Standing on this very spot, Robbins (1993) also poses a fair question: “for whom was the city once more public than now; for workers, women, lesbians and gay men, in other words, for the differences, minorities and the poor?”

Today, the main problematic aspect of the arguments for a better public space is that they have been articulated through narratives of loss and reclamation (Iveson, 2007). The result of this narrative loss has been a degree of false romanticization of historic public spaces (Madanipour, 2010), which becomes “a phantom1 , never actually realized in history but haunting our frameworks for understanding the present” (Iveson, 2007). In parallel, Berman (2012) also criticises as a timeless fact that public spaces have been a stage for the common people as subjects. The emergence of public space – Agora – as the place of appearance and inclusion into the political life dates back to the Hellenistic period of Western history and the emergence of the conditions of the first known democracy in the world; Athenian (Direct) Democracy. In other words, Agora became important enough to be shaped by the conditions of democracy. Agora, literally meaning, “the gathering place” was the centre of both socio-political and commercial life. Surrounded by commercial stoas and other administrative, cultural and religious structures, Agora was alive with people meeting, moving, talking and even just being present perpetually (Thompson, 1954). Yet, like the direct democracy itself, Agora was not a pluralist one. Although direct democracy has seen as a participatory model for all citizens regarding decision-making process (Urbinati, 2006), it is also often criticized as excluding women, slaves, old, children and foreign people from citizenship rights (Raaflaub, Ober, and Wallace, 2007). In other words, the tools and the rights to create the public sphere were in the hands of the young free Greek males who were allowed to be involved in political life in the public space of the Agora as the dominant publics (Martin, 2013; Mitchell, 2003). In fact, beginning from Agora of the ancient Greeks, public spaces have been an ideal place to learn both how to rule and how to obey (Berman, 2012), but not exactly to be included in the public sphere and to be heard for the common people and here we claim that it is the paradox of public

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

138

space. As Berman (2012) emphasizes, although Athenian Agora has often seen as a place where people could feel “at home” like Socrates, yet it was also a place that could kill him. In other words, the first democratic space of the world and so the democracy appears as a dream to be imagined, and this has been the romance of public space. Today, this romance becomes evident considering rising social struggles for public spaces especially following the crisis of 2007-2008. Various Occupy Movements might be seen as the signals of a paradigm shift in also urban planning that recently defined as “occupy urbanism” (Pak, 2016). In this very point, we should emphasise the paradoxical nature of the act of occupation of a public space turning back to its theoretically idealised definitions; “if they are open and accessible for all in practice, then why it is called as “occupy”2 , or should we understand the meaning of the act in the language like it is used on a toilet door writing “occupied” but meaning “in use” temporarily?” This question brings us further to investigate “paradox of political representation” developed by Pitkin (1967) and Runciman (2007). Here, it is important to highlight first that there is more than one type of political representation; it can be actualized in many different forms such as voting, joining a political party, signing a petition and other spatial campaigning tools like demonstrations and meetings. However, in this study, “paradox of political representation” has been tried to be interpreted only in terms of social production of public spaces considering also the importance and the necessity of physical representation of humans themselves. As the existing governmental system of modern states of today, what mainly distinguishes “representative democracy” from the “direct democracy” is the representative governments that had the effect of making it materially impracticable for the people to play a part in government and apparently even to be assembled. In fact, political representation has only been associated with the system of the election today (Manin, 1997).

Etymologically, the word representation derives from the Latin verb “repraesentare” of Roman law, yet, it was not an equivalent term used meaning “acting or speaking on behalf of someone else” as in the modern sense today. According to Skinner (2005), this basic meaning of the verb “repraesentare” was mainly “re-presenting something,” which means “bringing something missing or absent back into the present” in the two main contexts of law and after art. Vieira and Runciman (2008) explain the paradox of political representation based on representative democracy; although it is hard to know how democracy can work in practice, it is a purely political idea; however, representation is inherently ambiguous and paradoxical as it implies a presence and the absence that comes from the need to be re-presented simultaneously. Questioning the paradox beyond its etymology and putting it in the centre of the understanding of democratic politics; Runciman (2007) notes that it allows for the idea of representation to be identified both with the view that representatives should take decisions on behalf of citizens, and citizens should issue their instructions to representatives. In this very point, Pitkin (1967) puts forward one possible solution to the paradoxical nature of representation as “non-objection criterion”; “the substance of the activity of representing seems to consist in promoting the interests of the represented, in a context where the latter is conceived as capable of action and judgment, but in such a way that one does not object to what is done in his name.” Here, “objection” allows a kind of “presence” for those who are represented. This presence is based on the ability to object to what is done on behalf of represented people. Thereby, “silence” appears as a kind of confirmation whereas the “objection” offers the “presence” for the represented. However, representation takes place when there is no objection to what someone does on behalf of someone else, and political representation starts to break down when the explicit objections are voiced (Runciman, 2007).

ITU A|Z • Vol 15 No 1 • March 2018 • M. Demir Kahraman, B. Pak, K. Scheerlinck

The word “occupy” comes from the Latin word “occupare” meaning “to seize” (Bell, 2012, p. 216). 2

139

Figure 1. Moments of (social) space.

One might think of that whether the representation and so the representative governments are so fragile to be broken down because of explicit objections or not. However, the paradox of public spaces and why they are eventually “occupied” becomes clearer. In other words, why political power, the state or namely the representatives as the dominant discourse seek to reorganize, demolish, privatize, control, close or even militarize public spaces by police force also becomes revealed. Here, we claim that in parallel to political representation itself, social production of public spaces is paradoxical as the places where the explicit objections are voiced potentially. This claim further requires reinterpretation of the production of space and operationalization of Lefebvre’s concept of heterotopia to better understand the paradox of public space. 3. Social production of heterotopias as public spaces As a philosopher and sociologist, Henri Lefebvre considered space as a social product that serves as a tool of thought and of action claiming that a given space is both instrument and subject of production and it is a tool of control and dominant praxis, namely the power (Lefebvre, 1991). He also called each contextualization of his famous triads as “moments of social space,” which might be imagined like points in time or like a model of an atom that electrons hanging around the core; each has a gravitational force through its direction (one more than the other) producing the space. Though the production of space requires a process, the power of each gravitational force of each electron (context) would change presumably in time.

It gets clear in this diagram that what is conceived is not the same with what is lived and perceived, and/or vice versa; they are different moments of production of space and in terms of practical life; it is nothing but a paradox. In other words, what Lefebvre drew is both the political and paradoxical nature of social space. In fact, he did not use the terms public and private. Instead, he preferred “social space” based on “social relations of production,” and this relation grounds on struggle and so the production of space appears as a matter of politics. In conceptualizing of triads, Lefebvre (1991) did not mention any historical periodization since they can be applied to any period of time. Here, he rather referred to the shifts between modes of production that each produces its own space to show the change of the triads in the capitalist mode of production; and it remarks the simplification of the Western history. First, there is the “Absolute Space”; advantageous fragments of natural space to settle, which are soon populated. It is also the civil and religious space of agrarian population produced by peasants but managed by others (priests, warriors, etc.). Later, the preservation of the notion of family and what is sacred led the foundation of the political state, which remarks the “Historical Space.” “Historical Space” reflects the rise of representational space including religious and political symbolism that destroyed the absolute space and created the space of accumulation of the abstracts. But the shift to the feudal mode of production built its own “Medieval Space” tracing the representational spaces of the preceding and medieval town dominated the countryside, so does the bourgeoisie. Hence, “labor fell prey to abstraction,” and capitalism realized its own relations of production, creating its own “Abstract Space.” “Abstract Space” functions as a commodity with its exchange-value as a set of things and materials that are formal and measurable; it is the instrumental space of the authority, which erase differences. It is not only about constructing more or increasing the exchange value, but it is also about destroying;

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

140

Figure 2. Shifts between modes of production and their space.

not only the built environment but also the preceding social relations and differences composed of memories and identities eliminating the representational space. Critically, Lefebvre (1991) imagined the “Urban Revolution,” which would lead humanity to change what they have into a better life, only with the production of differential space and that would be realized by the dissolution of abstract space. Herein, “Differential Space” gathers differences towards the homogeneity of the abstract space of capitalism through appropriation and autogestion (Lefebvre, 2003; Harvey, 2014). There are two important distinctions between abstract and differential space. First, production of abstract space is the production of exchange value, but production of differential space implies a shift of use value over exchange value. Second, abstract space operates by signs and codes, which are attributed to them, whereas differential space operates by experiences and appropriations. In fact, societies that produce differential space reflect their self-representation in it, which is a creative and political process. Thereby, abstract space appears as a unity of lived and conceived spaces implying representations of space that coded by professionals and politicians while differential space appears as a unity of lived and perceived spaces implying spatial practice that appropriated by its users which would generate heterogeneous spaces and relations. As Butler (2012) emphasises, Lefebvre (2001) used the terms autogestion and appropriation to define social struggles as political resistance that transformed from abstract concerns and demands into concrete attempts to produce new spaces in saying; “each time a social group refuses to accept

Figure 3. Opposite characteristics of abstract and differential spaces.

and forces itself to master its conditions of existence, autogestion is occurring”. Herein, appropriation of space appears as the modification of a given space to serve the needs and possibilities of a group who appropriates it, and it requires the notion of property but in the sense of possession. Although Lefebvre mentioned that a site, a square or a street could be an appropriated space, it is often a structure: a monument or building. Since he developed the term referring the right to inhabitation; “right to the city” and “right to difference”; in this understanding, appropriation of space occurs in the form of squatting. In this point, Shields (1999) draws particular attention to that Lefebvre suggested squatting; slums, favelas, barrios, ranchos as re-appropriation of space; it is the birth of the tradition of “occupying” key spatial sites and buildings as a means of protest. Lefebvre (2001) gives the examples of squatting because, for him, differential space is the socialist space meaning the end of private property and domination of space by the power. Under the real conditions of capitalism, Harvey (2012) contributes at the most level to the literature considering the influential occupy movements around the world as the examples of

ITU A|Z • Vol 15 No 1 • March 2018 • M. Demir Kahraman, B. Pak, K. Scheerlinck

141

Figure 4. Production of heterotopia and isotopia. The concept of heterotopia is a medical term that describes a phenomenon occurring in an unusual place or a spatial displacement of normal tissue, which does not affect the development or organism as a whole (Sohn, 2008:41; Lax, 1997:115). It is introduced to the social sciences and urban theory for the very first time by French philosopher Michel Foucault, developed in “Des Escapes Autres” (1967) translated to in English as “Of Other Spaces” first in 1986 and as “Different Spaces” in 1998 by Hurley, but he never mentioned this concept in any of his writings (Hetherington, 1997:42, Saldanha, 2008:2082). Lefebvre’s discussion on heterotopia engages Foucault; in contrast to Foucault’s randomness, Lefebvre envisaged heterotopias from a political and historical perspective (Smith in Lefebvre, 2003: xii). 3

Figure 5. An urban context in Lefevbrian approach.

reclaiming the city and so the public space. Harvey refers Lefebvre’s vision of “right to the city” and mentions his concept of heterotopia as a clue to continue the search for understanding the social production and the paradox of public space. For Lefebvre (2003), social production of a heterotopia corresponds to the condition of anomie since he noted, “anomic groups construct heterotopic spaces, which are eventually reclaimed by the dominant praxis.” Here, by anomic groups, Lefebvre obviously referred to the counter publics in opposition to the dominant publics. Thereby, while Harvey (2012) considers occupy movements as the social production of heterotopias, we further comment that it is the spatialisation of the struggles between counter publics and dominant publics. Lefebvre (1991) pictured the concept of heterotopia3 in a historical formulation of “marginality” as “differentiation.” He categorized three “topias”

as the conceptual keys to explain the dissolution of abstract space: isotopia, heterotopia, and utopia. Here, utopia appears a non-place, but it seeks a place of its own; it is everywhere and nowhere hanging in the air in an urban context, embedded in the idea of monumentality. It is real and fiction, present and absent, yet, it is not a realized heterotopia. Further, monumentality is the fullness of a space beyond its material boundaries; it is plurality without contradictions so it is here and there within a differential and contradictive reality of urban space as a dream on the unity of differences, which can be assumed as his ideal, imaginary, revolutionary and permanent thinking of socialist space (Lefebvre, 2003). In this approach, isotopia also appears as identical places in neighboring order whereas heterotopia appears as the other place both excluded and interwoven in distant order. In this sense, isotopia refers to sameness as being identical, homologous and analogous; alongside them, there are different places as heterotopias (Lefebvre, 2003). In other words, during the crisis of capitalism and democracy, what is left from the dissolution of abstract space is isotopia, what is produced for a temporary period is heterotopia, and what is imagined is plainly utopia. Since competitive capitalism desirous of everything at the same time it produces abstract space, which is characterized by contradictions; quantity vs. quality, global vs. local, use value vs. exchange value (Lefebvre, 1991). This production is not stable, and these contradictions cannot be totally resolved, hence, abstract space cannot achieve full domination, and paradoxically this produces the differential space against itself (Harvey, 2014). To be more precise, apparently, isotopia corresponds to the “abstract space” while heterotopia corresponds to “differential space,” and three moments of space end up with a dialectic tension between them in the capitalist mode of production. Today, in the uneven conditions of capitalism and representative democracy, occupy movements appear as the examples of social production of heterotopic spaces that occur during

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

142

periods of crisis of domination when the counter publics appear capable of representing themselves and expressing their objections, yet, eventually reclaimed by the dominant publics. Developing the concept of heterotopia in relation with isotopia, Lefebvre gave us the chance to understand an urban context as a whole. From Lefebvrian approach, it appears that public spaces also might become heterotopias for a period of time; in other words, an urban context is composed of either isotopias or heterotopias during the dissolution of abstract space and the crisis of domination. Here, what Lefebvre focused on is also the production of space as a standardized commodity and its homogenous consideration in the capitalist mode of production. His critiques on the production of abstract/isotopic spaces overlap with the rising concerns after the 1970s, which argue that expected heterogeneity of public spaces has been violated by the capitalist and neoliberal practices more than ever in history. Further, like Harvey (2012), it is valuable to see his consideration of the social production of heterotopias (differential spaces) in the form of an act of occupation as the only way left to the differences (counter publics) today to make themselves visible politically. Also, it is significant to see that the struggles between dominant and counter publics overlap with the dialectic tension between isotopia (abstract space) and heterotopia (differential space). In fact, this study claims that what Lefebvre pictures is exactly the explanation of the social production of isotopias by the dominant publics and heterotopias by the counter publics producing their own representation through space. In brief, as the spaces of political representation, heterotopias are produced by the actual presence of counter publics and remark the plurality, yet they disappear just after the appropriation of the physical space ends and/or when suppressed by the dominant publics to make it end. From this approach, social production of heterotopic spaces overlaps exactly with the emergence of public spaces as the places of ideal enabling

to appear in the material human world in the sense of inclusion into the political life. To be more precise, as Cenzatti (2008) emphasizes heterotopias appear as the spaces of confrontation, which are produced through the social struggles between different publics. Here, it is revealed that the social production of a heterotopia is the manifestational realization of an ideal public space and the dissolution its paradox for only a temporary period of time when Lefebvre’s two moments of space (lived and perceived) gain enough force to appropriate the third (conceived). 4. Conclusion Above, we have discussed the contradictions between theoretically idealized definitions of public spaces and practical realities together with rising critiques on these contradictions. Integrating knowledge from the social sciences, we have developed this discussion further and revealed the paradox of public space. Out of this discussion, it appears that as the physical setting in which the public sphere has been produced and as the centres of everyday life, public spaces have been exclusionary as a representational tool of power. Since being public is a matter of inclusion in the political life and decision-making process, social production of public spaces have always been in the hands of a privileged spectrum of the societies who were allowed to shape the public opinion and rule as the dominant publics. Further, since neither domination nor publicness is possible in isolation and since power requires the justifiable obedience, public spaces full of crowds who are the objects of domination, appear as paradoxical to the extent until objects become subjects. In fact, this argument is directly related to the paradox of political representation. The simultaneous need for the absence and presence of other publics for the continuity of the representation of the dominant is explained by the paradox of political representation. Representative democracy system indicates the same paradox because representatives need the presence of people to be elected; however, they need the absence of people to represent

ITU A|Z • Vol 15 No 1 • March 2018 • M. Demir Kahraman, B. Pak, K. Scheerlinck

143

them simultaneously. Representation starts to break down when the explicit objections are voiced, in other words, when people prove their direct representation ability (publicness) giving the message that either they do not need representatives or elected representatives no longer represent them. The only solution to the continuity of the domination in some kind of balance is silence. Otherwise, the dominant confront with the possibility of losing control. This point remarks a greater problem that representative democracy as the existing governmental system of modern states of today itself appears paradoxical since non-objection contradicts with the essence of both the concepts of publicness and democracy. Besides, it proves that the paradox of public space is a timeless fact since although developing a comprehensive understanding of democracy has taken hundreds of years of human history; overall social mechanism remains the same without the plurality. Since the power is not stable, the counters of today might become the dominants of tomorrow. In accordance, the changing relations between multi-public drive the transformation of the public sphere and its spatial and semantical reflections on public spaces, because the moments of social space are also not stable and social production of space grounds on these relations. Thereby, an abstract space cannot be achieved full domination, and that creates differential space through appropriation and autogestion of counter publics, which remarks a crisis. In the capitalist mode of production, dissolution of abstract space appears as the dissolution of the domination and obedience. Herein, production of heterotopias as public spaces (differential spaces) relies on the human capacity to become the subjects of political action and discourse. However, since this capacity remains temporal being suppressed for the continuity of the domination, and since heterotopias of differentiation is eventually reclaimed by the power, what remains after a heterotopia disappears is the paradox of public space and political representation.

Today, from the M-15 in Puerta del Sol Square – Madrid to the Egyptian Revolution in Tahrir Square – Cairo and from the OWS in Zuccotti Park – New York to Occupy Gezi in Taksim Republican Square – Istanbul, occupy movements can be exemplified as the social production of heterotopias as the spaces of confrontation that temporarily occur during periods of crisis of domination. They are the manifestational realization of ideal public spaces and the dissolution the paradox for only a temporary period of time when Lefebvre’s two moments of space (lived and perceived) gain enough force to appropriate the third (conceived). As long as heterotopias are reclaimed by the power turning back into the abstract spaces of domination and as long as the permanent thinking of pluralist democracy is not achieved, the unity of differences will remain as a dream to be imagined; the utopia. Again paradoxically, it might be said that domination requires the production of heterotopias to prove and corroborate itself reclaiming them. Yet, as professionals from the domain of spatial sciences, our awareness of the struggling relations between multi-publics and their spatial or semantical reflections might also be useful to dissolve the paradox of public spaces providing the opportunity to confront the publics as the base to achieve ideal public spaces and so the pluralism. References Arendt, H. (1998). The Human Condition. 2nd ed. London: The University of Chicago Press. Bell, R. (2012). Public Space and Its Disconnects. In: Shiffman, R. et al. (eds.) Beyond Zuccotti Park: Freedom of Assembly and the Occupation of Public Space. Oakland: New Village Press, 214-236. Berman, M. (2012). The Romance of Public Space. In: Shiffman, R. et al. (eds.) Beyond Zuccotti Park: Freedom of Assembly and the Occupation of Public Space. Oakland: New Village Press, 197-207. Butler, C. (2012). Henri Lefebvre: Spatial Politics, Everyday Life and the Right to the City. USA and Canada: Routledge.

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

144

Carr, S., Francis, M., Rivlin, L.G. and Stone, A. M. (1992). Public Space. New York: Cambridge University Press. Cenzatti, M. (2008). Heterotopias of Difference. In: Dehaene, M. and De Cauter, L. (ed.) Heterotopia and the City: Public Space in a Post-civil Society, 1st ed. London: Routledge. Demir Kahraman, M., Pak, B., Scheerlinck, K. (2016). Making the Heterotopia: Gezi Park Occupy Movement. International Conference Contested Cities - Working Paper Series IV – 5A. Madrid: Universidad Autónoma de Madrid, 140-161. Demir Kahraman, M., Türkoğlu, H. (2017). Public; Up in the Air or in the Ordinal Scale. Journal of Planning – UCTEA Publications, 27(2), 141-151. Foucault, M. (1986). Of Other Spaces. Trans. J. Miskowiec, Diacritics. 16(1), 22-27. Foucault, M. (1989). The order of Things: An Archaeology of the Human Sciences. London: Routledge. Foucault, M. (1998). Different Spaces. In: Faubion, J. D. (ed.) Trans. R. Hurley et al. Aesthetics, Method and Epistemology. New York: New Press. Fraser, N. (1990). Rethinking the Public Sphere: A Contribution to the Cri- tique of Actually Existing Democracy. Social Text, 25/26, 56-80. Gehl, J. (1987). Life Between Buildings: Using Public Spaces. New York: Van Nostrand Reinhold. Habermas, J. (1974). The Public Sphere: An Encyclopedia Article (1964). New German Critique, 1974(3), 49-55. Habermas, J. (1991). The Structural Transformation of the Public Sphere. Cambridge, Mass.: MIT Press. Harvey, D. (2012). Rebel Cities: From the Right to the City to the Urban Revolution. Brooklyn – New York: Verso Books. Harvey, D. (2014). Seventeen Contradictions and the End of Capitalism. Profile Books: Kindle Edition. Hetherington, K. (1997). The Badlands of Modernity: Heterotopia and Social Ordering. London: Routledge. Iveson, K. (2007). Publics and the City. Wiley: Kindle Edition. Journal of the Society of Architectural Historians, 13(4), 9-14. Lax, S. (1997). Heterotopia from a

Biological and Medical Point of View. In: Ritter, R. and Knaller-Vlay, B. (eds.) Other Spaces. Graz, Austria: Haus der Architektur. Lefebvre, H. (1991). The Production of Space. Trans. Nicholson-Smith, D. Oxford: Blackwell Publishing. Lefebvre, H. (2001). Comments on a New State Form. Trans. Johnson, V. and N. Brenner, N. Antipode, 33(5), 769–782. Lefebvre, H. (2003). Urban Revolution. Trans. Bononno, R. Minneapolis: University of Minnesota Press. Lippmann, W. (1925). Phantom Public: A Sequel to “Public Opinion”. New York: Harcourt, Brace. Madanipour, A. (2010). Whose public space?: International Case Studies in Urban Design and Development. Abingdon, Oxon: Routledge, Taylor and Francis. Manin, B. (1997). The Principles of Representative Government. New York: Cambridge University Press. Martin, R. (2013). Public and Common(s). Places Journal, Retrieved from https://placesjournal.org/article/public-and-commons/ Mitchell, D. (2003). Right to the City: Social Justice and the Fight for Public Space. New York: Guilford Press. Pak, B. (2016). Strategies and Tools for Enabling Bottom-up Practices in Architecture and Urban Design Studios. Knowledge Cultures Journal, Special Issue (Learning, technologies, and time in the age of global neoliberal capitalism), 1-12. Pitkin, H. (1967). The Concept of Representation. Berkeley: University of California Press. Pitkin, H. (1968). Commentary: The Paradox of Representation. NOMOS X: Representation. R. J. Pennock and W. J. Chapman (eds.) New York: Atherton Press, 38-42. Raaflaub, K. A., Ober, J. and Wallace, R. W. (2007). Origins of Democracy in Ancient Greece. Berkeley: University of California Press. Robbins, B. (1993). Phantom Public Sphere. Minneapolis: University of Minnesota Press. Runciman, D. (2007). The Paradox of Political Representation. The Journal of Political Philosophy, 15(1), 93-114. Saldanha, A. (2008). Heterotopia

ITU A|Z • Vol 15 No 1 • March 2018 • M. Demir Kahraman, B. Pak, K. Scheerlinck

145

and Structuralism. Environment and Planning A, 40(9), 2080-2096. Shields, R. (1999). Lefebvre, Love and Struggle: Spatial Dialectics. London: Routledge. Sohn, H. (2008). Heterotopia: Anamnesis of a Medical Term. In: Dehaene, M. and De Cauter, L. (eds.) Heterotopia and the City: Public Space in a Post-civil Society, London: Routledge, 41-50. Soja, E. (1996). Thirdspace: Expanding the Geographical Imagination. London: Blackwell Publishing.

Thompson, A. H. (1954). Agora at Athens and the Greek Market Place. Tibbalds, F. (1991). Making People-Friendly Towns: Improving the Public Environment in Towns and Environment. London: Longman Publishing. Urbinati, N. (2006). Representative Democracy: Principles and Genealogy. London: The University of Chicago Press. Vieira, B. M. and Runciman, D. (2008). Representation. Cambridge: Polity Press.

Production of heterotopias as public spaces and paradox of political representation: A Lefebvrian approach

ITU A|Z • Vol 15 No 1 • March 2018 • 147-161

Accessibility of transfer centers with different transportation modes for disabled individuals

Berfu GÜLEY GÖREN1, Lale BERKÖZ2 1 berfuguley@gmail.com • Department of Urban and Regional Planning, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey 2 lberkoz@gmail.com • Department of Urban and Regional Planning, Faculty of Architecture, Istanbul Technical University, Istanbul, Turkey

doi: 10.5505/itujfa.2018.37267

Received: September 2017 • Final Acceptance: January 2017

Abstract Urban mobility of disabled people has crucial importance for integrated society. Disabled people have barriers in built environment. Thus, they cannot access urban spaces easily. Urban spaces are crucial for participation to public life for all. Identifying the barriers of mobility of disabled people and developing suggestions to eliminate deficiencies are necessary. In this study, universal design and accessibility standards have been integrated to show mobility of disabled people in urban spaces. Analysing table has been developed by AHP method to calculate the accessibility of disabled people. In case study, five transfer centers of Istanbul, which are in historical and central business area, have been analyzed and their scores have been calculated. By the analysis, current situation have been showed up and implementations that need to be done have been identified. The findings are going to help to local authority for accessibility implementations. Keywords Accessibility, Disability, Disabled people, Universal design, Urban design.

148

1. Introduction Disability is an important phenomenon of the contemporary world. In the process of globalization, urbanization is accompanied by a rapid increase in socio-economic activities and a large increase in population. Medical developments have been effective in prolonging human life, and the disability that comes with age has also increased in parallel with the disabled population. As stated in the United Nations’ and World Health Organization’s reports, 15% of the world’s population is disabled. According to the results of 2002 Turkey Disability Survey conducted in cooperation with the Office of Disability Administration of Turkey and the State Institute of Statistics, the proportion of disabled population in Turkey is 12.29% (TUIK 2002). In this case, the population of disabled people in Turkey is approximately 9 million, in İstanbul is 460 thousand. In addition to this population, relatives also face difficulties in their daily lives due to barriers, which deprive the disabled people of living an independent life. According to the TUIK datas, for 2013, the household is 3,7 in Turkey. Thus, approximately 33 million people face with difficulties in urban life and they are prevented from connecting with urban life and areas due to the lack of accessibility for all in the country. As it is stated in the Universal Declaration of Human Rights (United Nations General Assembly, 1948) every human being is equal in their respective rights to lead their lives as they see fit. Freedom of mobility is every individual’s right, regardless of people’s social, economic and physical characteristics. Every individual should be able to use services on equal conditions in the city. Public spaces are the most important elements of the city, where individuals come together to develop urban culture and engage in social interaction (Madanipour 1996). Public spaces that have great importance for integration of the society, public transportation and walking areas must be accessible for all. In literary studies, the accessibility of urban areas and universal design

have been two major topics. Yet, there is no evident study that correlates the two concepts of universal design principles and the accessibility of urban areas for disabled people. In this study, the concept of universal design and the criteria of accessibility have been anticipated to provide an independent life for all. The aim of this study is to create a new concept on the integration of universal design principles and the criteria of accessibility for disabled people in urban areas. Therewith this new composition would be included to the disability research literature. Furthermore, in this study a new matrix to calculate the accessibility score of transfer centers has been developed. The transfer centers of Istanbul and their relation to the public transportation have been examined in the direction of universal design principles and accessibility criteria. Assessment of accessibility of 5 transfer centers in Istanbul has been made by scores, which have been given by 10 disabled citizens of Istanbul, 4 of them are wheelchair users and orthopedicly disabled, 5 of them are visually impaired and 1 of them is hearing impaired. The accessibility scores of 5 transfer centers have been compared and their positive and negative qualities for accessibility has been exhibited. At the end of the study, to ensure full accessibility for all, urban design criteria have been determined. The use of the matrix, which has been prepared in the context of the study in urban design implementations by local governments is thought to be useful for the disabled to be active in urban areas. The principles of universal design are going to be at the core of the study that revolve-around seven principles, which envisage the importance of accessibility. The selected urban design elements, which have been inquired by 10 disabled citizens using the AHP method and have been calculated accordingly. 2. Literature review This section includes a literature research of the concept of universal design and accessibility that are the base of the study.

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

149

2.1. Accessibility of disabled people in urban areas In 2005, Independent Living Institute stated that: “Independent living is a philosophy, a way of looking at disability and society, and a worldwide movement of disabled people who work for self-determination, self-respect and equal opportunities”. According to the philosophy, the city can be described as a socio-physical space, in which individuals feel themselves as a member of the society they live in, since it is continuously renewed and productive. In the foreword of Jan Gehl’s book “Cities For People” (2010), Richard Rogers describes one’s right of accessibility in urban areas as follows: ‘Cities can be read as the books. The streets, the squares and the parks are the grammar of the city. These are structures that allow the city to have many activities in the city from quiet activities to loud and crowded activities. Everyone has the right to access open spaces just as they have the right to access clean water’ (Gehl 2010). Disabled people have stated that they are facing many architectural barriers. Due to those barriers, the disabled one is hindered from transportation, urban areas and possible employment.(Berube 1981). The right to the city, which is Lefebvre’s concept, indicates that every citizen may assert his/her own rights on the city they live in (Lefebvre, 1968). In this sense, the right of every citizen is to take place in the city and to use the city as an active member. The right to the city is an opportunity to eliminate inequalities in social life and to use the urban areas. In this respect, in the paper, urban design principles, which enable the right to the city, have been predicted that will provide accessibility for all. In this respect, recently, accessibility criteria should be applied in urban design with taking the relation and interaction between citizens and urban areas into consideration. Accessibility is a way to access housing, shopping, theater, parks and workplaces (CEAPAT 1996). Various projects are being carried out to alter the residential areas, public open spaces, educational areas, hospitals and transportation vehicles to include the disabled (NJSCC 2007; Gilman 2007; Igri,

2004). The way of creating the accessible urban areas is to take all possible users, including children, elderly people, adults and disabled into account. Accessible spaces can be defined as spaces without any architectural barriers. These spaces have the design or building features that provide accessibilty and promote mobility for all (Alonso 2002). An accessible design in urban areas allows disabled people to demonstrate their capabilities to play a vital role in the community. Many wheelchair users, visually impaired and hearing impaired individuals are involved in the public transportation system as key subjects of restricted mobility. Creating accessible urban areas and public transportation is crucial for independent urban mobility of disabled people. Accessible public transportation plays an enormous role in creating an inclusive society. Accessing the public transportation varies greatly around the world. Western Europe has some good examples of accessible public transportation, but serious issues remain untouched in some regions, particularly in Eastern Europe (International Disability Rights Monitor 2007; Steinfeld and Maisel 2012). During the past 15 years, researchers have adopted social ecological models to explain physical activity, which emphasize the importance of the built environment (Mcleroy et al. 1988; Sallis et al. 2006; Holle et al. 2014). Walkability reflects the built environments’ convenience for primarily transportation walking (Holle et al. 2014). While there is some new evidence that supportive attributes of communities’ physical environment can be associated with being more active (Ball et.al. 2001, Frank and Engelke 2001; Leslie et al. 2006). To clarify how built environment factors can infulence participation in physical activity, there is a need to identify and to document objectively, specific attributes of the communities’ environment that may be influential (Sallis et al.1998; Sallis and Owen 2002, Leslie et al. 2006). Increasing accessibility both on public transportation and in walkability is vital to ensure full accessibility for disabled people in urban areas. For this

Accessibility of transfer centers with different transportation modes for disabled individuals

150

reason, the matrix that has been prepared, includes the urban design criteria, which are necessary to ensure full accessibility in public transportation and in walkability. 2.2. Universal design in urban areas The concept of universal design arose from the disability rights movement, which began in the 1960s (Center, 2000). The concept of universal design aims to bring the disabled into the society by ensuring equal opportunity for all and eliminating discrimination based on disability (Steinfeld and Maisel 2012). The concept of universal design aims to create an environment, in which all users may realize all kinds of urban experiences by ensuring that every individual living in the city has access to all the urban areas with equal opportunities. Universal design is the art and practice of design to accommodate the widest variety and number of people throughout their life spans (Salmen 2011). In 1985, Ronald Mace defined universal design as; “The design of products and environments to be usable by all people, to the greatest extent possible, without the need for adaptation or specialized design” (Mace, 1985). The principle is to ensure that the environment and products are used without any other customization without regard to age, gender, ability and competence (Mace, 1991). Universal design increases the potential for developing a better quality of life for a wide range of individuals (Russell 1999; Stineman et al. 2003; Steinfeld and Maisel 2012). It provides benefits not only to people with functional limitations but also to society as a whole (Danford and Maurer 2005; Steinfeld and Maisel 2012). Universal design supports people in being more self-reliant and socially engaged. It helps change social stereotypes of disability and aging. Steinfeld and Maisel define that universal design is a process that enables and empowers a diverse population by improving human performance, health and wellness, and social participation. Universal design makes life easier, healthier, and friendlier. This process involves continuous improvement, based on the resourc-

es available, toward the ultimate goal of full inclusion (Steinfeld and Maisel 2012). Universal design includes principles that prevent the differentiation of disabilities. The concept that the World Health Organization supports and adopts takes every individual into account within the design regardless of its physical, social, economic characteristics. Universal design principles are aesthetic and functional solutions that are made without hinging not only upon the needs of specific individuals ,but as constructions that serve the public as a collective presence. A similar concept to the universal design is ‘design for all’. The term specifically is used in Europe (Steinfeld and Maisel 2012). In 2008, Design for All Foundation defined the term as design for human diversity, social inclusion, and equality (Design for All Europe, 2008). Design for All Foundation defines it as the “intervention on environments, products, and services with the aim that everyone, including future generations, regardless of age, gender, capabilities, or cultural background, can enjoy participating in the construction of our society, with equal opportunities participating in economic, social, cultural, recreational, and entertainment activities while also being able to access, use, and understand whatever part of the environment with as much independence as possible” (Design for All Foundation, 2015). Another similar concept is inclusive design. It is used particularly in the United Kingdom. In 2005, British Standards Institute defined the term as “the design of mainstream products and/or services that is accessible to, and usable by, as many people as reasonably possible . . . without the need for special adaptation or specialized design” (British Standards Institute, 2005). The Center for Universal Design conducted a research and demonstration project from 1994 to 1997, which was titled Studies to Further the Development of Universal Design, funded by the U.S. Department of Education’s National Institute on Disability and Rehabilitation Research (NIDRR). In this project, a set of universal design guidelines were developed.

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

151

The main objectives of the universal design concept are to improve the quality of life and to design environmentally and human-made systems to the specifications such as the user’s measurements, capabilities and constraints (Kroemer et al. 2001). The user-centered approach focuses on human diversity and is at the center of human factors and ergonomics discipline (Looze and Pikaar 2006). The principles of universal design aim to ensure that every citizen in the city enjoys equal access to all kinds of urban services. Universal design principles are found in a wide range of product scale to urban scale. Basic universal design principles applied to all design disciplines, including those that focused on built environments, products, and communications (Story et al. 1998; Center for Universal Design 2000a; Mueller 1997; Story 2011). In this study, the principles of the concept are explained in relation to the urban areas. Within the philosophy of universal design, seven principles have been identified as: equitable use, flexibility in use, simple and intuitive use, perceptible information, tolerance for error, low physical effort, size and space for approach and use as seen in Table 1. With these principles, the concept of universal design is more comprehensive than the concept of accessible design and barrier-free design. Universal design practices in urban areas make cities flexible and usable to provide utility for all types of urban users regardless of age, gender, proficiency and status. These solutions are required for the disabled as city users to live without encountering any barriers. One of the principles of universal design is “equitable use” which indicates each individual can use and access the urban areas with equal opportunities. A following principle is “flexibility in use” that is the design of urban areas and the use of transfer centers that are used to access urban areas, which are suitable for every individual with flexible designs. “Simple and intuitive use” is another principle, which means that every individual of the urban area is perceptible with simple stimuli and intuition regardless of their sufficiency status. The following principle is titled

“perceptible information”. This principle indicates that the stimuli and directions in the urban areas are organized in such a way that each individual can perceive, regardless of their limitations of perception. Another principle that universal design is based upon is named “tolerance for error”. It hinges upon the neccessity to take measures to reduce any risk of accidents that can occur in the urban areas. A following principle named “low physical effort” refers to the arrangement to minimize the use of physical force in mobility within the urban area. The last of these principles is named “size and space for approach and use”. This principle points out that the urban areas need to be organized in appropriate dimensions to provide an easy use for each individual in the urban areas. The integration of universal design principles and accessibility criteria will make urban mobility possible for every individual. Designing for the disabled is about making buildings and urban areas accessible and usable by people with disabilities. Universal design is about making urban areas safe and convenient for all their users, including disabled people. Furthermore, it means that the products, where the designs are universaly accommodating, that they cater conveniently for all their users. (Goldsmith 2000). To eliminate the marginalization of disabled people, accessibility criteria and universal design concepts should be used together to create accessible urban areas. 2.3. Integration of accessibility and universal design in urban areas Planning and constructing accessible urban areas have been the most ignored subjects for centuries. After the World War II, the disabled rights movement turned up claiming right on equal participitation on life all over the globe (Flesicher, 2001). Even though, most of the European countries have released technical accessibility standards and accordingly established anti-discriminination acts. Following these accomplishments, the awareness about accessibility has become an important topic. Unfortunately anthipathy and ignorance of the disabled in the community is still a wide spread attitude (Krpata 2012).

Accessibility of transfer centers with different transportation modes for disabled individuals

152

Table 1. The principles of universal design (Connell et al. 1997).

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

153

Since the last quarter of the 20th century, a great deal of efforts have been devoted to create accessible built environment (Steinfeld and Maisel 2012). In 1990, The Americans with Disabilities Act (ADA) draw great attention to the concept of accessibility and awareness on disability. Many different studies and legal regulations have been conducted on disability in several countries such as; the United Kingdom, Canada, Australia and New Zealand (Gleeson 1997). Turkey also enacted a law named the Disabilities Act in 2005 and is mandated to provide accessibility in public spaces. Nevertheless, it seems that accessibility is not fully achieved in practice in the country. In all these legal arrangements, accessibility is emphasized, but no emphasis is placed on the concept of universal design. There is a difference between universal design and accessibility. Accessibility is a function of compliance with regulations or criteria that establish minimum level of a design that is necessary to accommodate the disabled. The concept of universal design has many different dimensions and it emphasizes appropriate design for all people. Universal design and accessibility must not be considered separately from each other. The universal design and accessibility criteria need to be implemented in order to ensure that urban areas, structures, urban furnishings and all kinds of products needed in the city are used by everyone on equal conditions. The most important urban open space elements that connect the urban areas and provide mobility to the urban user are the streets. The arrangements to be made for this reason must start from the streets. In urban open spaces, signboards, urban furnitures, roads, ramps on the pavements, pedestrian crossings, all kinds of urban users should be considered and the new developments have to meet those usersâ&#x20AC;&#x2122; needs. Urban areas influence the individualâ&#x20AC;&#x2122;s participation in urban life. The relationship between the urban environment, the individual and society is a complex and comprehensive relationship. Universal design principles and

accessibility criteria are both crucial to ensure the continuity of these relationships. While integration of universal design and accessibility can be defined as communication process on the basis of functionality between user and environment, it can be defined as a means of strengthening the links between the environment, the individual and the community, and the urban physically. The integration of the concept of accessibility needs to be applied in practice in the seven principles of universal design in order to ensure the mobility of each individual within the city in all urban areas. The citizen whose needs cannot be met, consequently cannot feel him/herself as a part of the particular community, thus the city. Urban areas can be seen as the most important areas that meet the social needs of urban users, are streets, parks, roads, schools, hospitals, shopping centers and entertainment areas. These areas are where the people can be an active member of the city. For this reason, while designing the urban areas one should consider all its usersâ&#x20AC;&#x2122; needs. With the increase in accessibility and the emphasis on design for all, more and more experimental and observational researches have been conducted. Thus, creating more accessible urban areas for all types of users and practical solutions within the community have been established. Various experts such as; urban planners, architects, sociologists and politicians are involved in this process, while solutions are being sought to ensure that the city is used by all types of urban users. In order to find and implement effective solutions, these actors should have knowledge in theory and practice about the relationship between human and environment. Actors should create a language among themselves when applying solutions they find. Today, unconsciousness, ignorance, inconsistency, lack of communication between actors and contradictions cause the city not to be used equally by citizens. 3. Methodology and research area In this study, the Analytical Hierarchy Process (AHP) method, which is a multi-criterion decision making method, has been used.

Accessibility of transfer centers with different transportation modes for disabled individuals

154

Figure 1. Generic hierarchic structure (Bhushan and Rai, 2004).

The AHP method, is developed by Thomas L. Saaty based on the experience gained by him, while directing projects in the US Arms Control and Disarmament Agency. The AHP method was developed as a reaction to enable the taking of complex decisions. The AHP method has been used in social studies as a qualitative analysis technique. The AHP method’s universal adoption is a new paradigm for decision-making process with its ease of implementation. In the method, the subjective evaluations are converted into numerical values and processed to rank each alternative on a numerical scale (Bhushan and Rai 2004). In the study an accessibility matrix has been designed to assess the accessibility of disabled people in transfer centers by AHP method. In the matrix, universal design principles, which are identified in 1997 by the University of North Carolina’s Universal Design Center, became the main criteria. These criteria are equitable use, flexibility in use, simple and intiutive use, perceptible information, tolerance for error, low physical effort, size and space for approach and use. The sub-criteria of each main criteria are determined by considering the accessibility criteria of disabled people, that have been found in the literature researches and the needs of the disabled in the urban areas. 4. Determination of the urban design criteria, which are important in accessibility for disabled people, case study: 5 major transfer centers Istanbul is the largest city in Turkey. Istanbul developed a multi-centric metropolitan or megalopolitan character. Accessibility scores of five major transfer centers of Istanbul have been calculated by the matrix. These centers are Taksim, Kadıköy, Mecidiyeköy, Eminönü and Beşiktaş.

İstanbul was a vigorous core-dominated metropolis until well into the 1950s, with a very limited suburban development in the periphery. In the 1960s, the majority of jobs were in the core of the city. Because of this, in the historic centre Eminönü prouded transportation system initially by boats, cars, metro and buses. In the 1970s, new spatial structure was emerging decentralized in relation to the central business district (CBD). Istanbul is an old city whose long history has produced an interesting spatial development. (Dökmeci & Berköz, 1994). The reason for the selection of these centers is that they are located in the Central Business District (CBD) of Istanbul and they are transfer centers for public transportation for individuals, who come from every corner of the city. Taksim, Mecidiyeköy and Kadıköy have become new centers in CBD of Istanbul in 1960s. Eminönü is located in historical center and transfer center. Mecidiyeköy and Besiktas are located on the axis of the linear central business area that have gained importance in the 1970s (Dökmeci & Berköz, 1994). Main criteria and sub-criteria are seen in Figure 1, in which AHP method has been used. The aim of the study is to calculate universal design and accessibility scores of five transfer centers of Istanbul. After the determination of the criteria, weights of criteria have been given by 10 disabled people, with whom the interviews have been conducted (Table 2). Weights of criteria indicate the relationships between all the criteria. The hierarchic structure in the pairwise comparison of criteria on a qualitative scale. 1 means equal level between two criteria, 3 means one of criteria is marginally stronger than the other criterion, 5 means one of criteria is stronger than the other criterion, 7 means one

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

155

Table 2. AHP method with weighted percentages of all sub-criteria and main criteria.

Figure 2. Format for pairwise comparisons.

of criteria is very stronger than the other criterion, 9 means one of criteria is extremely stronger than the other criterion (Figure 2).

Universal design principles, which are the main critera of the matrix, have been compared pairwise by the disabled people. For example, “equitable use” criterion outweighs “flexibility in use” criterion and “size and space for approach and use” outweighs “simple an intuitive use” and “perceptible information” and “tolerance for error”, outweighs “low physical effort”. By taking all the scores of the disabled

Accessibility of transfer centers with different transportation modes for disabled individuals

156

people into consideration, the mean of the pairwise compairsons have been calculated (Table 3). For sub-criteria hierarchic comparison has been done by the same way. The comparison has been done for each criterion. The relations between all the criteria have been calculated by this method. All the hierarchic scores have been converted to percentages (Table 4 and Table 5). The same processes have been done for sub-criteria hierarchy. All the convertion of the AHP hierarchy scores have been applied for all criteria. The accessibility scores of the transfer centers are obtained as seen in Table 6. The weighting of the main criteria and the sub-criteria are determined by taking the average of the weights, which have been given by the disabled people. All the sub-criteria of the matrix has been scored from 1 to 4 by the disabled people for five major transfer centers in Istanbul. For the scaling of the sub-criteria the Likert scale has been used in the study. The Likert-type or frequency scales use fixed choice response formats and are designed to measure attitudes or opinions (Bowling, 1997; Burns, & Grove, 1997). The description of scaling in the Likert scale has been given as follows; 1: not absolutely appropriate 2: not appropriate 3: eligible 4: fully appropriate mean scores. The averages of the given scores by participants for all the main and sub– criteria have been calculated then the final score table for each center has been obtained. As a result of the tables, universal design and accessibility scores of Istanbul’s five major transfer centers, have been calculated and compared with one-another. In the comparison between the centers, the percentage of the scores obtained from each main criteria for each center have been calculated. Furthermore, the arithmetic averages of these percentages have been taken and the accessibility scores of 5 major transfer centers of Istanbul have been determined. 5. Findings In the analysis, if all the criteria are met, a transfer center would score

Table 3. Hierarchic comparison of main criteria.

Table 4. Convertion to percentage of hierarchic comparison of main criterion.

Table 5. Hierarchy of main criteria.

as 4. While, in a case where the transfer center would not meet any of the criteria it would score as 1. Depending on this resolution, Taksim has scored

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

157

Table 6. Five transfer centers’ universal design score results.

Table 7. Transfer center-based covarage ratios of universal design main criteria.

Table 8. Distribution of 7 main criteria for all transfer centers.

the highest accessibility score, which is 2.79, and Beşiktaş has scored the lowest accessibility score, which is 1.28. As a result, none of the five transfer centers are fully accessible for disabled people. Renovation works that have been carried out in Mecidiyeköy and Eminönü. They have low accessibility scores because accessibility design criteria have not been provided yet. The reason why the accessibility score in Taksim is higher than other centers is that the renovation work has been completed at this transfer center. In Taksim; it is expected that the accessibility score will be higher than 2.79 despite the fact that the regulation work has already been completed. This score reveals the incompleteness of the implementation in this transfer center. There has not been any renovatory or regulatory work done in Besiktas. The reason that this centre has ranked the lowest score can be illustrated by its

lack of a strong connection of its different modes of transportation and the spatial distance between the aforementioned modes of transportation. As a result of the analysis; the principle of “equitable use” has given a result of 87% in Taksim, 80% in Kadıköy, 58% in Beşiktaş and 50% in Eminönü. When the average of these five centers have been realized considering the equitable use principle, the value would equate to 65.2%. The principle of “equitable use” has the greatest proportional value among the other principles in terms of fulfilling the criteria (Table 7 and Table 8). “Tolerance for error”, “low physical effort” and “size and space for approach and use” main criteria have similar average values equate to a range of 50%. ‘Flexibility in use’ main criterion has scored 42.4% average value. ‘Perceptible information’ main criterion has scored 38%, ‘simple and intuitive use’ main criterion has scored 38.6%; which is the lowest score. These two main criteria include particularly the urban design criteria, which meet the needs of people with sensory disabilities. This low scoring is the result of taking only the orthopedicly disabled citizens into consideration while conducting the matrix. When the criteria of the transfer centers with the highest and lowest rates are considered; When in Taksim, the subcriteria of “equitable use” provides %87, while the sub-criteria of “flexibility in use” and “perceptible information” provide a scoring of 41%. According to this matrix’s results, Taksim has scored a high score considering the main criteria of ‘equitable use’ and ‘size and space for approach and use’, yet has scored a low score in ‘flexibility in use’ and ‘simple and intuitive use’ and “perceptible information” main criteria, despite the renovation efforts in the transfer center. In Kadıköy, the sub-criteria of ‘tolerance for error’ main criterion has provided 85% while sub-criteria of “simple and intuitive use” main criterion have provided 25%. According to Table 6; Beşiktaş, has the lowest accessibility score (1,28) in five transfer centers, according to Table 8; 58% of the sub-criteria of the ‘eq-

Accessibility of transfer centers with different transportation modes for disabled individuals

158

uitable use’ main criterion have been provided, while 32% of the sub-criteria of the “perceptible information” main criterion have been provided. In Eminönü, 51% of the sub-criteria of the ‘equitable use’ main criterion have provided, while 29% of the sub-criteria of the ‘perceptible information’ main criterion have been provided 29%. In Mecidiyeköy, 53% of the subcriteria of the ‘tolerance for error’ main criterion have been provided, while 29% of the sub criteria of ‘flexibility in use’ main criterion have been provided. The 5 transfer centers of Istanbul are crucial for the mobility of the urban citizens. If an accessible transfer center project is to be designed in Istanbul, it is crucial to use such a matrix as the integration of universal design and accessibility criteria. The current situation of the centers shows that Taksim largely contributes to urban mobility of disabled people with ‘equitable use’, ‘simple and intuitive use’ and ‘size and space for approach and use’ criteria; Kadıköy largely contributes to urban mobility of disabled people with ‘flexibility in use’, ‘perceptible information’, ‘tolerance for error’ and ‘low physical effort’ criteria (Table 9). In Figure 3, the circle diameters are determined by the universal design and accessibility scores of the centers. As seen above, Taksim, Kadıköy, Beşiktaş, Eminönü and Mecidiyeköy rank from the highest to the lowest. In Figure 4, there are images from the 5 transfer centers. In the analysis, ongoing renewal works in Eminönü and the ongoing metro construction in Mecidiyeköy cause the lowest accessibility scores for these transfer centers. Renewal works that have not been completed yet affect accessibility negatively. The findings of the study, which assess the situation of urban design elements by AHP method, that are important in the accessibility for disabled people in the central areas in Istanbul, suggest that urban design elements are inadequate in transfer centers and that quality is far lower than the required standard. Taksim has the highest accessibility score, all the sub-criteria of ‘equitable

Table 9. Center-based proportion of universal design main criteria.

Figure 3. Five transfer centers of Istanbul.

Figure 4. Five transfer centers of Istanbul.

use’ have been implemented but “the height of railings of stairs and ramps formed at two different levels as 60 cm and 90 cm” criterion has been ranked inefficient. All the sub-criteria of ‘flexibility in use’ have been implemented but “fold-

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

159

able seats are at the stops”, “stops located at the same level with vehicle”, “public toilets are suitable for disabled people” and “stops’ sizes are sufficient for a wheelchair user” criteria have been ranked inefficient. All the sub-criteria of ‘simple and intuitive use’ have been implemented but ‘the warning signs on pedestrian crossings are indicated by light and phosphorus colours”, “staircase and ramp edges are highlighted with contrasting colours” criteria have been ranked as low quality. Two of subcriteria of ‘perceptible information’ main criterion have not been implemented, which are “guiding & routing panels prepared with large pinpoint letters & Braille” and “information & orientation panels prepared with Braille”. Three sub-criteria have been ranked as low quality, which are “indication of the surroundings of urban furniture with remarkable coluors & materials”, “information & voice information in the directional panels”, “the height of the information and guidance panels is 120 cm to be seen by wheelchair users”. All the sub-criteria of ‘tolerance for error’ have been implemented but “the floor of the covers on the paths must be maximum in 1,25 cm level difference”, “sudden constrictions or extensions at the cross section of the paths” and “positioning of urban furnitures (required a minimum of 2 m height to prevent head injures)” have been ranked as low quality. All the sub-criteria of ‘low physical effort’ have been implemented but “embossed material, not higher than 5 mm on the surface of paths”, “suitable resting areas in certain periods (60100m)” and “corresponding ramps on the opposite sidewalks” have been ranked in low quality. All the sub-criteria of ‘size and space for approach and use’ main criterion have been implemented but “paths’ and ramps’ width must be at least 90 cm”, “maximum height of pavement must be max.15 cm”, “ramp gradient must be max. 5%”, “positioning of stimuli/ markers at appropriate points to provide guidance” and “steps on boarding platform are in same level” have been ranked as low quality.

As a result, ‘simple and intuitive use’ and ‘perceptible information’ criteria, which have scored the lowest in all the transfer centers, primarily have to be improved. Furthermore, the other main criteria should be raised up to the top level as %100. 6. Conclusions In this study, the quality of urban design elements and the conformity of the accessibility criteria of five major transfer centers of İstanbul, that have connections between the different modes of transportation in the central business area of Istanbul, have been assessed. The state of the urban design elements, which are important in the accessibility for all. Therefore those have been determined in the framework of the seven principles of universal design, have been evaluated by 10 disabled individuals. The results have been calculated by the AHP method. Urban design should create democratic and equal habitats for all people. To achieve accessibility, universal design concept and its principles must be adopted in urban areas. Accessible urban areas provide equal protection, equal opportunities for all and create social justice and human-centered urban environments. To increase independence of mobility of disabled people in urban areas, all urban design implementations should be based on accessibility criteria. The design of transfer centers, with the adoption of universal design principles, is very important for the independent mobility of disabled people. Sub-criteria, which have been considered for planning and design studies should be provided according to the universal design principles. All kinds of design implementations are important to ensure accessibility and mobility in urban areas. All the urban areas should be designed in accordance with the needs of disabled people to create fully accessible urban areas. Accessibility score calculation matrix, which has been prepared in the study, will be available to local governments to calculate the accessibility of transfer centers’ designs. Full accessibility will be ensured when all of the design criteria can be applied to the implementations.

Accessibility of transfer centers with different transportation modes for disabled individuals

160

References Alonso, F. (2002). The Benefits of Building Barrier- Free: A Contingent Valuation Of Accessibility As An Attribute Of Housing, European Journal of Housing Policy 2(1), 25-44. doi: 10.1080/14616710110120577. Ball, K., Timperio, A.F., Crawford, D.A. (2006). Understanding Environmental Influences on Nutrition and Physical Activity Behaviors: Where Should We Look and What Should We Count ?, International Journal of Behavior Nutrition Physical Activity, 3 (33). doi: 10.1186/1479-5868-3-33. Berube, B. (1981). Barrier-Free Design-Making the Environment Accessible to the Disabled, Canadian Medical Association Journal 124, 68-78. Bhushan, N., Rai, K. (2004). Strategic Decision Making Applying the Analytic Hierarchy Process. Springer Publishing. Bowling, A. (1997). Research Methods in Health. Buckingham: Open University Press. British Standards Institute, (2005). Design Management Systems, Managing Inclusive Design. Burns, N., & Grove, S. K. (1997). The Practice of Nursing Research Conduct, Critique, & Utilization. Philadelphia: W.B. Saunders and Co. Center, A.E. (2000). Designing for the 21st Century, An International Conference on Universal Design. <http:// www.designfor21st.org>, Accessed 27 December 2017. Center for Universal Design, North Carolina State University (2000). Universal Design Exemplars, Raleigh, N.C.: Center for Universal Design North Carolina State University. Center for Universal Design, North Carolina State University (1997). The Principles of Universal Design, http:// www.ncsu.edu/www/ncsu/design/ sod5/cud/about_ud/udprinciplestext. htm, Accessed 01 February 2017. Centro de Referancia Estatal de Autonomia Personal y Ayudas Tecnicas (CEAPAT), (1996). Concepto Europeo de Accesibilidad [European Concept for Accessibility]” Madrid: IMSERSO. Connell, B. Jones, R., Mace, M.L., Mueller, R.L., Mullick, J.L., Ostroff, A., Sanford, E. J., et al. (1997). The Principles of Universal Design, Version 2.0, Raleigh, N.C.: Center for Universal

Design North Carolina State University. Design for All Europe (2012). Design for All. www.designforalleurope. org/Design-for-All/, Accessed 12 April 2016. Design for All Foundation (2015). What Is Design for All?, www.designforall.org/en/dfa/dfa.php, Accessed 01 January 2017. Dökmeci, V., Berköz, L., Transformation of Istanbul from a Monocentric to a Polycentric City, European Planning Studies,Vol.2, No.2, pp.189-20, 1994. Fleischer, D., Zames, F. (2001). The Disability Rights Movement: From Charity to Confrontation. Philadelphia: Temple University Press. Frank, L., Engelke, P. (2001). The Built Environment and Human Activity Patterns: Exploring the Impacts of Urban Form on Public Health, Journal of Planning Literature 16 (2), 202–218. doi: 10.1177/08854120122093339 Gehl, J., (2010). Cities for People” Washington, DC: Island Press. Gilman, S. (2007). Including the Child with Special Needs: Learning from Reggio Emilia, Theory into Practice, 46(1), 23-31. doi: 10.1080/00405840709336545. Goldsmith, S., (2000). Universal Design, Oxford: Reed Educational and Professional Publishing. Holle, V.V., Cauwenberg, J.V., Dyck, D.V., Deforche, B., Weghe, N.V., Bourdeaudhuij, I.D. (2014). Relationships Between Neighborhood Walkability and Older Adults’ Physcial Activity: Results from the Belgian Environmental Physcial Activity Study in Seniors (BEPAS Seniors), International Journal of Behavioral Nutrition and Physcial Activity, 11:110. doi: 10.1186/ s12966-014-0110-3. Igri, L., (2004). Social Context, Special Needs / Disability / Developmental Disorders and Educational Inclusion, Hrvatska Revija Za Rehabili- Tacijska Istrazivanja, 40(2), 151-164. Indipendent Living Institute. (2005), https://www.independentliving.org, Accessed 27 December 2017. International Disability Rights Monitor (IDRM) (2004). Regional Report of the Americas, Chicago: Center for International Rehabilitation, Washington,

ITU A|Z • Vol 15 No 1 • March 2018 • B. Güley Gören, L. Berköz

161

DC.: Center for Internantion Rehabilitation . International Organization for Standardization (2001). Accessibility and sability of the built environment”. https://www.iso.org/committee/291991.html, Accessed 22 May 2017. Kroemer, K.H.E., Kroemer, H.B., Kroemer-Elbert, E. (2001). Ergonomics: How to Design For Ease and Efficiency, Upper Sandle River, NJ: Prentice Hall. Krpata, R., (2012). Accessible Public Transport: Vienna City Tourism, Best Practice in Accessible Tourism, Inclusion, Disability, Ageing Population and Tourism, Canada: Channel View Publications. Lefebvre, H. (1968). Le Droit à la ville [The right to the city] (2nd ed.). Paris, France: Anthropos. Leslie, E., Coffee, N., Frank, L., Owen, N., Bauman, A., Hugo, G., (2006). Walkability of Local Communities: Using Geographic Information Sytstems to Objectivly Assess Relevant Environmental Attributes, Health Place , 13 (1), 111-122. Looze, M., Pikaar, R. (2006). Editorial: Meeting Diversity in Ergonomics, Applied Ergonomics, 37 (4)I 389-390. doi: 10.1016/j.apergo.2006.04.002. Mace, R. (1985). Universal Design: Barrier Free Environments for Everyone, Los Angeles: Designers West. Mcleroy, K, Bibeau, D., Steckler, A., Glanz, K. (1988). An Ecological Perspective on Health Promotion Programs, Health Education Quarterly, 15, 351-377. Mueller, J. L. (1997). Case Studies on

Universal Design, Raleigh, N.C.: Center for Universal Design North Carolina State University. New Jersey Schools Construction Corporation (NJSCC ), 2007. 21st Century Schools: Design Manual, New Jersey. Sallis, J., Bauman, A., and Pratt, M. (1998). Environmental and Policy Interventions to Promote Physical Activity, American Journal of Preventive Medicine, 15 (4), 379–397. Sallis, J.F., Owen, N. (2002). Ecological Models of Health Behavior, Health Behavior and Health Education Theory Research and Practice, San Francisco: Jossey-Bass. Salmen, J. P. S. (2011). U.S. Accessibility Codes and Standards: Challenges For Universal Design, Universal Design Handbook, USA: McGraw Hill. Steinfeld, E., Maisel, J. (2012). Universal Design, Designing Inclusive Environments, Hoboken, New Jersey: John Wiley & Sons, Inc. Story, M.F., Mueller, J.L., Mace, R.L. (1998). “The Universal Design File: Designing for People of All Ages and Abilities”, Raleigh, N.C.: Center for Universal Design North Carolina State University. Story, M. F. (2011). The Principles of Universal Design: Universal Design Handbook, USA: McGraw Hill. United Nations General Assembly (1948), http://www.un.org/en/universal-declaration-human-rights/. Accessed 22 May 2017. Wolfgang, F. E. P., Korydon, H.S. (2011). Universal Design Handbook, USA: McGraw Hill.

Accessibility of transfer centers with different transportation modes for disabled individuals

ITU A|Z • Vol 15 No 1 • March 2018 • 163-175

An evaluation on immaterialisation phenomenon in religious spaces of architecture

Ümit ARPACIOĞLU1, Mustafa ÖZGÜNLER2 1 umitarpacioglu@gmail.com • Department of Arcitecture, Faculty of Architecture, Mimar Sinan Fine Arts University, İstanbul, Turkey 2 ozgunlerm@gmail.com • Department of Arcitecture, Faculty of Architecture, Mimar Sinan Fine Arts University, İstanbul, Turkey

doi: 10.5505/itujfa.2018.85520

Received: April 2017• Final Acceptance: December 2017

Abstract This study relates to the subject, but with a specific focus on the material and the conceptual approach of monumental religious architecture to the use of material. Material is originally formless; in a constant quest to find expression. Especially with industrialization, material, that started to be considered independently of construction process of a particular artifact has been displaced (deterritorialization) and deformed. The ‘Architecture-Reality” relation lies in the interaction between the material and its representation, which gives the material its expression. This point of view will lead us to understand how the use of material finds expression and help us define the ways in which religion uses the material to shape the space. This the study will elaborate on the relation of ‘Religion-Architecture-Reality’ with respect to the concept of dematerialisation. To define what can be a common expression for all three elements, we must openly look at singular examples; with a perspective independent of geography, culture and time. It is quite difficult to notice and point out how these concepts emerges in spaces. Laying out the different setups and the perceptional differences created within the setup can certainly enable a better definition of the relevant methods. This study evaluates the religious spaces related to the two widespread religions in terms of material use and religious expression; in an attempt to draw attention to the two contemporary concepts we have borrowed from art. Keywords Immtaterialization, Dematerialization, Arts, Islamic architecture, Technology.

164

1. Introduction The ‘Architecture-Reality” relation lies in the interaction between the material and its representation, which gives the material its expression. Material is originally formless; in a constant quest to find expression. As form is not inherent in the material, but is imposed on it by the mind of the producer, it represents the idea of one who renders it (Gönül, 2014). Especially with industrialization, material, that started to be considered independently of construction process of a particular artifact has been displaced (deterritorialization) and deformed. The resettlement (reterritorialization) and the re-acquisition of a certain form and expression of matterial is realised by design process of a particular artifact. In this sense, material is beeing displaced (deterritorialized) while it is resettled (reterritorialized) in the production-consumption relations of the modern world (reterritorialization) (Koçyiğit, 2007). For centuries, designers have tried to define the bond between material and perception, and pondered about creating a product that exceeds the limits of perception. This subject in architecture is a wide-ranging issue that can be studied with respect to many dimensions. This study relates to the subject, but with a specific focus on the material and the conceptual approach of monumental religious architecture to the use of material. To highlight once again, a universally valid result can only be obtained through the perspective of many experts. The ‘Religion-Architecture-Reality’ relation is found at the intersection of the conscious – or unconscious – expression of religious thought. The relation between “religion, architecture, reality” can be found at the interface of the conscious-or unconscious expressions of religious thoughts.To define what can be a common expression for all three elements, we must openly look at singular examples; with a perspective outside the context of geography, culture and time. This point of view will lead us to understand how the use of material finds expression and will help us to define the ways that the religious

understanding/approach uses the material to shape the space. This part of the study will elaborate on the relation of ‘Religion-Architecture-Reality’ with respect to the concept of dematerialisation. It is quite difficult to notice and point out how these concepts emerge in spaces. Laying out the different setup and the perceptional differences created within the setup can certainly enable a better definition of the relevant methods. The structural design and spatial organization of a religious space can be analysed through countless perspectives and systematics. This paper develops an analytical/a different approach for the evaluation of religious space in terms of use of material and its religious expression. The aim of this study is to emphasize the two contemporary concepts that are quoted from art through the examples of most common two religions: Islam and Christianity. Immaterialisation is based on material objects that are subjected to human mind as George Berkeley (16851753) claims, “matter does not exist and only minds and thoughts exist. In our minds material beings becomes perceivable and phenomenal. The concept of immaterialization is used only for 20th century works, products of 20th century architectural approaches. However, the ideological siblings of the same works are found deep in history, a surprising fact that shows us how old these concepts actually are. This study also demonstrates how the concept of immaterialization can be and has been emerged to give direction to human thought. 2. The temporal superiority of religious architecture (timelessness) The Religious space has a character of becoming a perceptual symbol that centralizes itself within its environment, in which its size determines the symbolic power. When examples of religious spaces – especially those that are designed for large-scale urban areas – are examined, it can be seen that these spaces are meant to achieve superiority against time. The designer contemplates on a religious space design; not only in terms of material,

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

165

technology, arts and planning, but also in an effort to carry it beyond time. The reasons why this behavior is so commonplace among all architects are part of a larger and multi-faceted issue. The search for being timelessness in religious space can be classified as; • Physical Timelessness in contrast to human life: The architect designs spaces with a desire to render them timeless, superior and long-lasting in comparison to the human life span. This aspiration places many responsibilities on the architect. In this manner, this architecture is different than industrial design. The design must preserve its expression, for those who have not been born yet. According to human spatial experience, continuity can be defined in two ways. In the first approach, the continuity define by perceived spaces depending on the human movement while in the second approach it is defined by the change and transformation of a space over time (history) (Koçyiğit 2002, 43-60). We are focusing on the second one in this study.The designer accepts much more accountability for a religious space, a space especially full of meanings. • Timelessness in arts and aesthetics: An architectural product can be preserved easily, depending on material and technology, but it also has to beat time in terms of aesthetic preferences and artistic quality of design. In other words, an architectural product must be designed with an innovative approach, representing long-term ideals and preferences of society. Today, with the advent of new material and technology, it is much faster and easier to build any structure. It is the perspective that makes the difference. It is the architect who uses a language that speaks to people of all times, who will not let the design lose artistic value after a hundred years. • Timelessness in shaping social life: When the architect is designing a space s/he seeks solutions that meet specific needs. However, as human lifestyle changes and develops through time, so do these needs and

expectations. The architect then is obliged not only to think about the structure of the monument, but also has to contemplate on future lifestyles. Human living transforms dramatically in a hundred – or two hundred – years. The question in the architect’s mind should be: What kind of function will this space provide for the people, a hundred years from now? At this point, an architect who designs a special mystical space such as a mosque, must definitely have a strong grasp of universal human needs, and a strong imagination. Interpreting the factors mentioned above require execution of important analyses for a religious space. Temporal superiority in the real sense can only be achieved when all these factors are considered. These kinds of eclectic mosques have a negative effect on the temporal superiority of the building; since the material is not as long-lasting as masonry. This mixing of less durable materials with more traditional and persistently enduring designs can have a perceptual effect such that there is a great chance that, a hundred years later, the visitors to these mosques will be confused, because their temporal perception will not recognize this eclectic feature. One of the predictions for the city of future, is the differentiation of its’ inhabitants’ perception with the context-free use of images and materials; “Rather than its continuity, the habitants of copy-paste city should experience the temporality of the built environment”(Postalcı, Kuruç Ada, & Özbek Eren, 2006). While Anatolia, and of course the whole Islamic geography, is full of examples that have achieved timelessness in all meanings, today’s mosques are rarely designed with an aspiration to survive through time. 3. Immaterialisation and dematerialisation phenomenon in religious architecture: Islam and Christianity Immaterialisation in architecture is an phenomenon that dematerialises by degreasing the perceptive visibility of the architectural structure and components or representing as materialistic

An evaluation on immaterialisation phenomenon in religious spaces of architecture

166

Figure 1(a-b). Pont-Neuf Bridge, Paris, 1985 (on the left) and Reichstag Parliament Building, Germany, 1995 (on the right) by Christo and Jeanne Claude (Hasol 1999)(Anon 2014c).

absence that produces unity in mind. According to Sayın (Sayın, 2016), dematerialisation is also a fact that produces body and soul unity through realistic/techtonic and materialistic differentiation affected with the world by transformation in architectural transparency. In this article immaterialisation phenomenon in architecture is determined under two categories in materiality. First, dematerialistic phnomenon of the material use dematerialising by degreasing the materialistic perception and bringing forth the architectural form by harmony in time. Second, immaterialisation phenomenon of the material use by illusions and simulations in time. Dematerialist phenomenon of the material use; In architecture, form is defined by the display of the material used and this adds an extra depth to the architectural shape. For example what we consider most about a natural stone is its weight and durability where glass is associated with transparency and fragility. These meanings of quality are developed through a thousand years’ experience. Creating the architectural structure by means of a homogeneous, even integrated approach by decreasing this experience based on perceptual qualities of the material, thus emphasizing the form and geometry, is called “Dematerialization”. This type of architecture brings form and geometry upfront and makes the structure dematerialize by making it partially artistic as a sculpture. The most distinct attribution of dematerialization is the use of covering; not to change the look of the structure but making the covered structure more recognizable architectually. One could take a

Figure 2. Steel construction parts used by S. Calavatra can be seen as an example of dematerialist use of the materials. The form created by the use of steel is more pronounced than the material itself (Anonim 2012).

look at the works of Christo and Jean Claude to get to a better comprehension of the state of dematerialization which leads the work to become an artistic, material-less sculpture. These two artists covered Pont-Neuf bridge, Paris, with polyamide fabric and ropes in 1985 and Reichstag Building, the old parliament building in Germany, with polypropylene fabric in June 1995. This juncture points out dematerialization as simplicity, skillfully achieved by “materiallessness” seen most especially in today’s visual aesthetics (Kansu, 2000)(Meiss, 1991). Figure 1 shows Pont-Neuf Bridge, Paris, 1985 (on the left) and Reichstag Parliament Building, Germany, 1995 (on the right) by Christo and Jeanne Claude (Hasol, 1999)(Anonim, 2014d). Figure 2 shows Steel construction parts used by S. Calavatra can be seen as an example of dematerialist use of the materials. The form created by the use of steel is more pronounced than the material itself (Anonim, 2012). Today, designers come up with products based mostly on form to diminish the perceptual effects of the

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

167

Figure 3. The mosque architecture in Islam reached its peak point in Anatolia integrating mysticism and design. It has a dematerialist approach in the use of materials, bringing form upfront(Anon 2014d).

Figure 4(a,b). Immtaterialization / Space relationship(Jodidio 2012)(Başbuğ 2000).

material. Designers should know the material and comprehend its meaning to the society very well to make use of dematerialization. It is only after this point that the material can transcend its perceptual quality; from the physical to the semantic. A century of structural codes and practices and what we perceive as steel is changed by a distinct form as can be seen in figure 2. In a way, dematerialization is acquired or achieved through a form beyond the standard use of the material. Figure 3 shows The Mosque Architecture in Islam reached its peak point in Anatolia integrating mysticism and design. It has a dematerialist approach in the use of materials, bringing form upfront(Anonim, 2014e). Immaterialist phenomenon of the material use ; Another use of material

in architecture is found in the concept of immaterialization, usually defined by illusions and simulations, where the virtual and the real collide. The roots of this concept lie in capturing the haste and spontaneity of life (L Wittgenstein, 1990). Architecture is an expression of what is thought to be real, but this expression is becoming more of a “vision expressing itself in a different way” than what we’ve known before. We can say that Immaterialization is the process of integrating the structure and environment in a flow of information (Dilekçi, 2000)(Başbuğ, 2000). Figure 4 shows Immtaterialization / Space relationship(Jodidio, 2012)(Başbuğ, 2000). We can examine dematerialization in two different ways as mentioned above. Religious architecture itself also used this technique to impress, direct and make its believers feel the religion’s power and understand its philosophy. The philosophy of Islam has always found life in new discourses in order to establish a common language with the people of different cultures. Although this common language is usually vaguely defined, or left undefined; we can actually grasp the nature of it by referring to its concrete products: sanctuaries. Islamic architecture refers to all relevant works around the ancient world; to name a few; in the Arabic, Andolous, Ottoman, North African, Middle Asian regions. It may be thought that these religious spaces are shaped by national values, and that the national culture has a dominant influence on architecture. Though the nation consept has not showed up until 18th centuary. It is sayable that culturel, communal, traditional values and climatic conditions, materials, constructional traditions of all these areas had effected the architecture. Nevertheless some concepts generating from religion of Islam’s necesities and ideas had also effected the architecture. As a result some similarities have arised in mosque places. However true this analysis is, this study will look past the national elements; rather it will focus on the influence of religious thought on the religious space. To truly understand Islamic architecture, one has to adopt

An evaluation on immaterialisation phenomenon in religious spaces of architecture

168

a well-adjusted abstract lens; looking beyond regional cultural elements, ethnicity or the locally available material. This study aims not to find explanations for all elements of Islamic architecture; but to find new points of view by looking into the fundamental ideas in Islamic thought, and of course how these ideas are demonstrated in Islamic architecture. Limiting this study to the abstract lens is done for two reasons. The first one is the obvious one: to isolate the object of study. Although the name of the study can refer to many aspects regarding the subject; the topic of interest is based on Islamic thought and the direct reflection of Islamic ideas on Islamic architecture. The second reason explains why the abstract is strikingly important regarding the object of study: Because Islamic architecture, like all monumental architecture, has tended to be conceptually higher than the national culture of the region. The aforementioned search for a “common language” found in the base of Islam’s expansion, the attempt to blend ideas with cultures has gained considerable power to Islam, thus to Islamic architecture. Due to centuries of cultural synthesis, it is possible to find vast differences in mosque design and structure, across the Islamic world (Hattstein & Delius, 2000). Understanding Islamic architecture and the way it shapes the environment requires looking through an abstract lens; independent from the geographical limitations, regional culture, local material, racial and ethnic differences. If one can adopt this abstract view of Islamic spaces, a rather interesting fact becomes visible: that an interaction with Islamic philosophy has tended to advance the architecture of a culture, in comparison with other regional or national cultures within the same geographical area. The main reason for Islam’s positive effect on architectural design is that Islam has survived throughout centuries by way of cultural synthesis, which, in architecture, simply means adaptive power(Cansever, 2012). Designers make use of specific techniques to render their design unique, epressive, memorable, immortal, or

monumental; a tendency which defines the unique style of the space, as well as the use of material and technology in the making of that space. An inquiry of these methods shows that the philosophy of Islam has influenced much of the character of Islamic architecture. But if religious philosophy has such a powerful and diffused influence on design and space, how can we isolate it from other factors of design? How can we really understand this phenomenon? Will the perspective we acquire give us the opportunity to analyse the religious spaces of the past and present? It is important at this point to view this issue with a holistic approach, elaborating on abstract concepts rather than individual examples. The ways in which Islamic thought shapes and reshapes architectural forms is an issue that can be discussed through countless lenses. This study primarily focuses on the use of material, the construction technologies involved and the relationship with the physical environment. It is likely that the discussion would revolve around many more important topics, if more and other factors were included for a wider perspective. However – and therefore – for the sake of a clear discussion this study will be confined to the above mentioned elements. As more studies approach the issue from various perspectives, it might hopefully become possible to make more comprehensive evaluations. It is another crucial point of departure that, while discussing Islam’s basic approach to spaces in terms of material use and technology; it is also necessary to discuss the differences between Islam and other forms of religious thought on the same issue. This is not to discriminate between belief systems, rather, the aim of this study is to understand the structural design forms that arise from differences in the philosophy and beliefs that make up these religions. • In Islamic architecture, the concept of Tawhid (unity) plays an important role in terms of style and material(Cansever, 2012). We must first elaborate on the idea of Tawhid (unity), an important aspect of Islamic belief, as well as Islamic archi-

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

169

Figure 5(a-b). Islamic architecture has tried to use material without exaggerated emphasis, and to preserve the material’s natural structure(Anon 2016).

tecture. The concept is used to create a holistic sense, to create a link between the space and the belief. In other words, the concept is constructed as space, and when visitors enter the space they are not only surrounded by design elements but also by the concept itself. This is a feature that strengthens the believer’s perception of Islam – making it possible for one to grasp the idea through the space, which, in return, makes it possible for the idea become more powerful; thus creating a cycle of growing belief, the main reason why sanctuaries are built in the first place. It encourages collective prayer and emphasizes a social structure in which all believers are equal. This ideal of unity is spatialized in the form of mosque architecture. All elements serve their function, without being placed in the foreground, a holistic perspective that reflects quest for perfection in the final product. This design approach is also an effort to support Islam’s approach to social stratification. An egalitarian approach can be found in many examples of mosque design, in many design elements from the distribution of light to material use and architectural planning. In order to define the concept of Tawhid (Unity) in spatial terms, one must already have a strong grasp of the philosophy of Islam. The designer then can make an interpretation, which also has to fall within the boundaries of the current conditions of material, technology and cultural atmosphere. The designer must be able to seed this concept into whatever culture and conditions the mosque will live in. In the hands of the right designer, the mosque becomes, or at least should become, and end product of Islam that is expressive of its ideals and supportive of its reach to all humanity.

The objective was not to discover the ideal upper structure, but to create the largest single uninterrupted space disturbed by as few vertical structural elements inside the main prayer hall as possible. In other words, the common denominator in early Ottoman mosques is not the form of the interior, but the nondirectional containment of the inner space by four walls. And despite the many influences throughout its evolution this characteristic feature of the Ottoman mosque architecture remains unchanged(Aptullah Kuran, 1986)(Abdullah Kuran, 1987). Architect Sinan’s works are some of the most important mosques in Islamic architecture. These are works that merge religious philosophy with the material (Hattstein & Delius, 2000). • Islamic architecture has a clear structure that is both silent and active. It adopts a humble and natural style. In examples of Islamic architecture, we see that nothing is unnecessary. There are, of course, exceptions to this general condition. However the plain, unexaggerated use of material and technology is more common. In addition, the mosque is a space that is constructed with available local resources, which makes it a humble synthesis of Islamic thought and the local culture. Another factor that accounts for the sense of humility in these spaces is that the emphatic and focussed sanctity attributed to most religious spaces is often missing for mosques. They are functional spaces that provide for public needs. Figure 5 shows Islamic architecture has tried to use material without exaggerated emphasis, and to preserve the material’s natural structure (Anonim, 2016). • Islamic architecture uses material as it is, neither denying nor over-em-

An evaluation on immaterialisation phenomenon in religious spaces of architecture

170

phasizing its natural qualities. The resources of material for Islamic architecture are various, due to the changing geographical area and climate. Therefore, a specific type of material cannot be said to characterize mosque design in the global sense. However, there are many similarities in the use of material. Knowing oneself and demonstrating one’s actual self are some of the key ideas behind Islamic thought. This ideal has contributed to the architecture in the form of humility and simplicity. The material is not attributed a different form than the one it already has. • Clarity of form is also a result of certain notions in Islam, such as deep consciousness, responsibility and mightiness. In a way the material and form have a harmonious relation. This harmony reflects Islam’s ideals of the self; one’s search for one’s whole self in pursuit of perfection and enlightenment. When we look at the meanings attributed to spaces, we see that the mosque emphasizes ideals of centralism and equality. The expression of these ideals is supported by the use of material, technology and light. Whereas in sanctuaries of other religions, separation of groups and a clear emphasis on clergy are quite common; these are also supported by the architectural features of the space. That is to say; two different spaces may be built by the same material using the same technology, and the message that is conveyed through the space may differ vastly. Figure 6 shows In Islamic and mosque architecture, the space is egalitarian. There’s no orientation towards one person or group, as there is in many other religions (Anonim, 2014a)(Anonim, 2014c). • In Islamic architecture, form and material are not attributed any sanctity. Iconization is denied in Islam. Therefore, holiness is not restricted to the space. The most important feature of monumental architecture is that it is loaded with meanings, of course as far as these meanings can be perceived. Therefore, the size of constructions is always parallel with the importance of the mes-

Figure 6. In Islamic and mosque architecture, the space is egalitarian. There’s no orientation towards one person or group, as there is in many other religions(Anon 2014a)(Anon 2014b).

sage that they are conveying. When the space setup is compared to that of Christianity, the predominant religion in Western cultures, Islam does not directly attribute any meaning or sanctity to the space itself. The space is expected to provide messages that are supportive of Islamic thought, but not to direct the visitor’s gaze or attention. This characteristic is parallel to some ideas in Islamic thought; such as self-awareness, social awareness, finding one’s own path to enlightenment without using it as a means for material success. “It is not the space that is holy” is the basic perception that, to a certain extent, sets the design elements free in mosque architecture. Thus, the mosque as a space is not a repetition of religious themes. Mosque architecture is of a changing and developing nature. The best examples of this characteristic can be found in 16th century Anatolian mosques. Regarding its view against deifying icons, its egalitarian approach and simplicity of materials used, it can be said that Islam has created many spaces closer to the dematerialist style. The use of material in Christianity on the other hand, is based on immaterialist style; positioning the altar in the structure, emphasizing the importance of the church for the society; class distinctions influenced the design. Although immaterialization and dematerialization are relatively new concepts in architecture and art, it is important to realize that these elements can nonetheless be seen in use in religious spaces built hundreds of years ago. Figure 7 shows Two domes; one covering an Islamic space (left)(Hattstein & Delius,

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

171

Figure 7(a-b). Two domes; one covering an Islamic space (left)(Hattstein and Delius 2000); the other covering a Christian space (right)(Anon 2014e).

Figure 8(a-b). Illustrates an Islamic dome (on the left) and another one which belongs to a Christian religious space (on the right) (Anon 2014e).

2000); the other covering a Christian space (right)(Anonim, 2014b). Image 8 illustrates an Islamic dome (on the left) and another one which belongs to a Christian religious space (on the right). Despite al the embellishments, dematerialist use of structural elements in Islam is clear; their structural function is evident, none of them outshine the other and each of them is “present” in the structure. In the other picture we see the dome as a gate to the heavens, which is an immaterialist use of a dome. The immaterialist style leads the observer to see what should be seen – by illustrating the message directly – whereas the dematerialist style leads one to perceive the uniformity of the structure, how parts sum up as a whole – by demonstrating the message through architectural elements. On the dematerialist side, we see the structural elements as they are; no visual illusions, ornaments don’t outshine the structural elements thus keeping the uniformity of inside and

outside. But in the immaterialist example it’s hard to recognize the structural elements, not even the space itself. Neither of these styles is superior or inferior to one another. These are just two different practices of religious philosophy and sociology, finding their own way to make their spaces impressive. In figure 9, Two different uses of elements regarding interior structure(Gülçubuk, 2014)(Ludovico, n.d.). There are big differences between Islam and Christianity regarding interior design. Islam makes an equal effort to reach perfection in internal and external spaces, thus creating an interior-exterior unity at mosques. Also the visual composition of structural elements is upfront regarding interior design. In western cultures, Christianity has adopted a space setup that makes use of painting arts, and this setup conveys clear messages, creating a didactic imagery or experience in the interior space by visuals. This is not an illustration that the visitor chooses to

An evaluation on immaterialisation phenomenon in religious spaces of architecture

172

Figure 9(a-b). Two different uses of elements regarding interior structure (Gülçubuk 2014) (Ludovico (n.d.)).

experience. This illustration or fiction has been installed to be delivered to him/her. The paintings of religious stories and personas, lead to a perception that the religious space may serve as a mediator between God and the rest of society rather than as a place of enlightenment. The decor and structure in many Christian religious spaces can also support or convey the hierarchical position of the clergy. In Islam no figurative references are used in structural design. The religious space is set up as a tool of enlightenment, in which the visitor can find his or her own self, catch a glimpse of the perfection of Allah, and sense the idea of unity (tawhid). As mentioned above the two most widespread religions and their relative paradigms have utilized different styles to make their religious spaces more impressive. Even if it would not be right to categorize the religious spaces only according to the use of material, this has been attempted for the discussion of a practical issue. The portrayal of differences between two methods of material use can give us a perspective and make way for new criticism of religious architecture. Through the perspective of what this study offers, today’s Islamic structures can be re-evaluated in terms of their ideological proximity to the thought system they belong to. More perspectives of a similar kind should definitely contribute to an ideal synthesis in Islamic architecture. When talking about improving religious architecture, the relevant art forms should not be forgotten. Since all religious thought has been supported by art, it is an un-

deniable fact that art gives power to the message that the religious space is trying to convey. An ideal synthesis in Islamic architecture would also be expected to bring stronger artistic expression in religious spaces. Art is a form of thought which leads and shapes society. All belief systems have adopted a unique artistic interpretation. This artistic approach is a reflection of the feelings that religion awakens in society, reciprocally giving strength to each other. Therefore arts and crafts related to religious philosophy have been a reference point for society, while at the same time being influenced by society itself. In this case, it is only natural for differences in religious philosophy to account for differences in artistic approaches. Islam is a belief system that is widely accepted by many societies and deeply embedded in the lifestyles of populations. For this reason it is no surprise that Islamic art, as a whole with all its parts, has many strengths and a unique character. This study is constrained to those characteristics of Islamic art that are manifested in mosque architecture(Katz, 1979). The illuminated colors support the style of the whole, which celebrates happiness, spreading joy and hope all around. Islamic philosophy emphasizes the importance of the individual’s enlightenment. The mosque’s artistic design supports this ideal of englightenment and search for the self. The motifs, as well as the use of colors and light, verify this principle in Islamic thought. All artistic elements in the mosque can be said to have a support-

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

173

Figure 10(a-b). Examples of Islamic art (Hattstein and Delius 2000)(Katz 1979).

Figure 11(a,b). The differing ways in which religious art has used material and techniques for illuminating the space(Hattstein and Delius 2000)(Anon 2014a).

ing effect on each other. Art does not individualize; it is of a collective nature, in congruence with the Islamic concept of tawhid, unity. This is also the reason why the signature or signi-

fier of the artist is not pronounced on the work(Cansever, 2012). In Islamic art, the tradition of portraying human misery is not found, in contrast to some other religions. In Islam, art refers to the holistic character of the space. As the mosque takes a functional role between God and its subject; any feature that could influence or intrude in this bond is prohibited. Islamic art, unlike forms of art that have developed within some other religious belief systems, does not assume a responsibility or an ideal to control or shape human behaviour. It is not concerned with conveying a direct message. In other words, there are no thought leading elements in the profane features of mosque architecture, but only those awakening feelings of wonderment and awe. Figure 10 shows Examples of Islamic art (Hattstein & Delius, 2000)(Katz, 1979). During the Ottoman period, the state organization included centres for architecture research and education, such as Ehl-i HÄąref and State Guild of Architects. These guilds have contributed greatly to the development of arts and crafts in the empire. This organized community of artisans had a privileged status within the state, a considerable authority on the practice of architecture and a supervisory role on the related practices. This placement of architects in the Ottoman Empire had created a culture in which art was supported by the state and the state

An evaluation on immaterialisation phenomenon in religious spaces of architecture

174

had worked its means to preserve the quality of mosque construction(Aydın, 2004). Figure 11 shows The differing ways in which religious art has used material and techniques for illuminating the space (Hattstein & Delius, 2000)(Anonim, 2014a). 4. Conclusion Islam has left a unique mark on all cultures it has influenced. We need to understand these marks, which have made it to today through cultural synthesis, to be able to understand them in today’s conditions. Mosque architecture, among these art forms, is an important space in the sense that it conveys social messages in addition to demonstrating Islamic philosophy to its visitors. Considering this heavy context in which mosques are designed and built, we also need to fully understand all those social values that have become synthesized Islamic art. Even as individuals, our spatial perception changes in time and in response to social change. Examples of this interactive change can be found in Ottoman architecture, especially of the 16th century, a period in which mosques were created to become symbols of Islamic thought. Today we still refer to 16th century works of Islamic religious architecture in Anatolia, and try to reproduce the style of design, unaware of the fact that our imitations fail to represent the philosophy of Islam. The fact that these products are promoted today is an important sign of Islam’s separation from its core philosophy. This is one of the striking effects of modernity on Islamic thought, sweeping away the universal ideals it is based on, and leaving its believers unable to engage in the culture of self-discovery, novelty and authenticity. A recent controversy in Islamic architecture is about the problems of representation inherent in mosques built with today’s technology but in the 16th century style. Addressing this phenomenon with a contextual focus on material use, the mentality behind building mosques with reinforced concrete and designing them to look like stone-masonry constructions can be defined as immaterialist, rather than

the dematerialist style to which they were originally and perhaps more properly connected . Today a space that is designed to serve Islam can ironically be of a structure that is incongruent with the philosophy of Islam. This contradiction tells us more about how people have come to interpret religion as a set of appropriate behaviors, and how this view of religion has affected views on practical matters. To reconsider the ways we use material and technology, we need to free ourselves from these fixed scenarios on what is religiously correct. It will only then be possible to speak of contemporary Islamic architecture, where we may start observing the emergence of new forms. Only then we can create spaces that have the power to carry Islamic thought to the future, build structures which reflect the true meaning of Islam in today’s language. References Anonim. (2012). SANTIAGO CALATRAVA. Retrieved from http:// www.calatrava.com/ Anonim. (2014a). 9. St Vitus Cathedral in Hradcany. Retrieved from http://www.docbrown.info/docspics/ europe/czech09.htm Anonim. (2014b). CHURCH CEILING DOME MURAL PROJECT; PART 1. Retrieved from http://tracyleestum. com/st-matthews-cathedral-ceilingdome-mural-project-part-1/ Anonim. (2014c). ERKEN DÖNEM OSMANLI MİMARİSİ. Retrieved from http://www.gateofturkey.com/ section/tr/616/7/kultur-ve-sanat-mimari-osmanli-donemi-mimarisi Anonim. (2014d). Jeanne-Claude und Christo biography. Retrieved from http://www.wikiartis.com/en/jeanneclaude-und-christo/ Anonim. (2014e). Sultan Ahmet Camii. Anonim. (2016). Bursa Ulu Camii. Aydın, A. (2004). 16. yüzyıl Osmanlı yapılarını etkileyen bâni-sanat-sanatçı ilişkisinin İstanbul camileri özelinde incelenmesi. Yıldız Teknik Üniversitesi, İstanbul. Başbuğ, E. (2000). Bilgisayar Teknolojisindeki Gelişmelerin Mimari Malzeme Teknolojisine Etkileri. Yıldız Teknik Üniversitesi.

ITU A|Z • Vol 15 No 1 • March 2018 • Ü. Arpacıoğlu, M. Özgünler

175

Cansever, T. (2012). İslam’da Şehir ve Mimari (7.Basım). İstanbul: Timaş Yayınları. Dilekçi, D. (2000). Elektronik Paradigmaya Geçiş ve İmmateryalite. Domus Dergisi, 4. Gönül, H. (2014). Mimarlıkta Formsuzluk Kavramı Üzerine Bir Araştırma. Yıldız Teknik Üniversitesi. Gülçubuk, M. (2014). Eşsiz Tavanlar. Retrieved from http://www.gonuldergisi.com/essiz-tavanlar-dr-mehmet-gulcubuk.html Hasol, D. (1999). Mimari izlenimler”. YEM Yayınları. İstanbul: YEM Yayınları. Hattstein, M., & Delius, P. (2000). Islam Art and Architecture. France: Könemann. Jodidio, courtesy of P. (2012). Egg of winds. Photo courtesy of Philip Jodidio. Retrieved from http://architecturalmoleskine.blogspot.com.tr/2010/06/ toyo-ito-tribute-to-winds.html Kansu, B. (2000). İmkanların Malzemesi. Domus Dergisi, 4, 72–73. Katz, J. (1979). Architecture As Symbol and Self-Identity. Morocco: The Aga

Khan Award For Architecture. Koçyiğit, R. G. (2007). Mimarlıkta Yersizleşme ve Yerin Yeniden Üretimi. Mimar Sinan Fine Arts University. Kuran, A. (1986). Mimar Sinan. İstanbul: Hürrüyet Vakfı. Kuran, A. (1987). Sinan: The Grand Old Master of Ottoman Architecture. İstanbul: Ada Kitapları. L Wittgenstein. (1990). The Duty of Genius. New York: Penguin Books. Ludovico. (n.d.). Contest of Heavenly and Earthly Love: Fight for the Palm. Meiss, P. Von. (1991). Elements of Architecture. New York: Spoon Press. Postalcı, E., Kuruç Ada, A., & Özbek Eren, İ. (2006). The New Urban Memory. Cities Between Integration and Disintegration, Opportunities and Challenges. 42nd International Society of City and Regional Planners (IsoCaRP) Congress. Sayın, T. (2016). Kartezyen Dualistik Düşünceden Ontolojik Çözünmeye: Mimarlıkta “Geçirgenliğin” Dönüşümü. In Mimarlıkta Işık Sorunsalı (pp. 22–55). İstanbul: MSGSÜ Yayınları.

An evaluation on immaterialisation phenomenon in religious spaces of architecture

Contributors Ümit ARPACIOĞLU Born in 1976, in Istanbul. He received his bachelor’s degree at Mimar Sinan Fine Arts University at 2001, on the same year he began his career as a research assistant in the Department of Architecture at MSFAU. He completed his master and PhD theses at the Institute of Science and Technology, Building Physics and Material Department. Since 2012, he continues his career as an assistant professor doctor at MSFAU. Miray BAŞ YILDIRIM Miray Baş Yıldırım practices as an architect in Melbourne, Australia. She has recently completed her PhD degree from Istanbul Technical University. Her research focuses on computational design, urban regeneration and participatory design. She currently holds a master’s degree in architecture from Middle East Technical University. Ayşe Lale BERKÖZ Ayşe Lale Berköz is full professor at Faculty of Architecture, Department of Urban and Regional Planning at Istanbul Technical University. Hugh David CLARKE Hugh Clarke is a British architect who was elected to the RIBA in 1984. His teaching practice seeks to draw upon his work as a professional architect and urbanist. Current interests include planning for changing identities in the contemporary city. Gülen ÇAĞDAŞ Gülen Çağdaş has graduated from Istanbul Technical University, Faculty of Architecture. She took her Ph.D. degree on Architectural Design from the same university in 1986. She has become Professor in 1997. She is still the Department Head of Informatics at the Institute of Science, Engineering and Technology, ITU. Pınar ÇALIŞIR ADEM Pınar ÇALIŞIR ADEM graduated from Karadeniz Technical University, Faculty of Architecture in 2009. In 2012,

she took her master degree in Bartlett School of Architecture. She is currently a Ph.D. student in ITU in Architectural Design Computing Program and continues her career as a lecturer at Yeditepe University. Ömer Halil ÇAVUŞOĞLU Ömer Halil Çavuşoğlu is a PhD Candidate at Istanbul Technical University. He is also working there as a research assistant at Architectural Design Computing Program. His main research area is Building Information Modeling (BIM). Ebru ÇUBUKÇU Ebru Cubukcu is Professor of City and Regional Planning Department in the Faculty of Architecture, Dokuz Eylul University, Izmir, TURKEY. She received her PhD from The Ohio State University, US. Her research interests focus on perception and cognition, virtual environments, walkability, post occupancy evaluation, and creativity in design education. Meriç DEMİR KAHRAMAN Meriç Demir Kahraman is a lecturer and holds a PhD from Istanbul Technical University - Faculty of Architecture, Urban and Regional Planning Department. She has an experience of five years as a Research & Teaching Assistant at ITU - Faculty of Architecture and a semester as an International Scholar at KU Leuven - Faculty of Architecture. Elif ENSARİ Elif received her B.Arch degree from METU and her M.Arch degree from Sci-ARC. She co-founded the architectural design studio Iyiofis in 2011 and the research practice Bits’n Bricks in 2016. Currently, she is a Phd candidate at Istanbul Technical University and University of Lisbon developing methods for measuring urban walkability. Berfu GÜLEY GÖREN Berfu Güley Gören is PhD Candidate at Department of Urban and Regional Planning at Istanbul Technical University. At the same time she is lecturer at Department of Architecture at Istanbul Arel University.

Orkan Zeynel GÜZELCİ Orkan Zeynel Güzelci is currently a research assistant at the Department of Interior Architecture & Environmental Design at IKU, and PhD student in Architectural Design Computing Program at ITU. He received his MSc (2012) degree in Architectural Design Program at ITU. During his graduate studies, he won various prizes in national architectural competitions and completed their construction drawings. Bilge KOBAŞ Bilge received her BArch degree from Istanbul Technical University (2008) and MSc degrees from ITU (2010) and AA (2012). She established her design studio Super Eight in 2014, and co-founded research practice Bits ‘n Bricks in 2016. She has been teaching design studios and sustainable design courses in several universities. Mustafa ÖZGÜNLER Professor Dr. Mustafa Özgünler is an architect. He worked between 19922009 as a research assistant in Faculty of Architecture at Istanbul Technical University. Since 2009, he has been working as an Assistant Prof. Dr. in Department of Environmental Control and Building Materials Technology at Mimar Sinan University of Fine Arts. He has been working about building materials, laboratory test methods, environmental control and fire security. Burak PAK Burak Pak is a Professor of Architectural Collaborative Design, Collective Spaces and Digital Media at KU Leuven Faculty of Architecture. He holds a PhD from ITU Faculty of Architecture. He is currently teaching design studio courses and running international and national research projects in Brussels and Ghent Campuses. Burkay PASİN Received his B.Arch degree in Architecture from METU (2000), his M. Arch degree in Architecture from DEU (2007), and his Ph.D. degree in Architecture from METU (2014). Currently works at Department of Architecture at the IUE. His major research interests

include modern architecture, gender and space, and architectural education. Dagmar REINHARDT Dr Dagmar Reinhardt is a principal of reinhardt_jung (www.reinhardtjung. de), a widely published architecture practice that received multiple awards for residential work. She teaches design and design research as the BAEnvs Program Director and leader of Digital Architecture Research, and develops robotic architecture applications driven by structural simulation and acoustic performance. Ceyda SARICA Ceyda Sarica is Urban Designer & City and Regional Planner at the Plan and Project Department of Karsiyaka Municipality, Izmir, TURKEY. She is a PhD candidate at the Graduate School of Natural and Applied Sciences, Dokuz Eylul University. Her research interests focus on post occupancy evaluation, spatial perception, environmental psychology. Kris SCHEERLINCK Dr. Kris Scheerlinck is Architect and Urban Designer and is currently a Professor at KU Leuven. He ran his own research and design practices in Ghent, Barcelona, London and New York while lecturing at various institutions and universities. He is Director of NYhub, a research and design platform in New York. He directs a research project on collective spaces, called “Streetscape Territories” and promotes related PhD projects. Sinan Mert ŞENER Born in İstanbul, in 1961. He started his architectural studies in İTÜ, in 1978 and earned his graduate degree in 1984. He completed his PhD. degree in 1994 and conducted his post-doctoral studies in Carnegie Mellon University, between 1998-1999. Şener, worked as a Research Assistant between 19881995; as an Assistant Professor between 1995-1997; as an Associate Professor between 1997-2011 and he has been working as a Professor since 2011 in İTÜ, Faculty of Architecture. He served as Dean’s Assistant between 2000-2008 and as Dean between 2012-2015.

Güzden VARİNLİOĞLU Received B.Arch. from METU (2001), M.F.A. (2003) and Ph.D. (2011) degrees from Bilkent University. Was a visiting scholar at ITU and UCLA (2011-14). Currently works as Assist. Prof. at the IUE. Research interests include computational design, digital heritage, and design education.

Şeymanur YILDIRIM Şeymanur Yıldırım is an architect at Doğuş İnşaat and PhD student at Istanbul Technical University (ITU), Architectural Design Computing Program. She has experience on Algorithm Aided BIM from professional practice and master thesis, which is written by her on this subject.