Portfolio - 2014 Examples of thoughts, concepts, solutions, and drawings by Scott Levine
Oracle - Big Data Discovery
Oracle
Big Data Discovery Senior UX Designer
Oracle - Big Data Discovery
Background Endeca to Oracle Big Data Discovery When I joined the Endeca team at Oracle, they were in a bit of a bind. They had a great technology, but their Information Discovery product was not selling like anyone thought it should. They were starting to feel pressure from Oracle to make some headway, and soon. When I arrived, the priorities for UX were to start to find a way to address the needs of a self service user (to remove the IT configuring layer), and to start understanding what it would take to bring a reasonable experience to mobile users. My first major responsibility was to come up with a vision for what the mobile experience could be. This redesign effort set in motion a series of events that lead to the creation of a brand new, category-defining product: The first holistic big data discovery application for Hadoop.
Oracle - Big Data Discovery
The Problem I joined right as the Endeca Information Discovery version 3.1 was to be released. It was a brand new redesign, and looked like this:
The technology was great, but this application of it didn’t serve anyone well. It was an IT configured “grey label� product built on a portal framework with a package of basic visualizations such as charts, tag clouds, maps, etc., but sold to a largely self service oriented market. To a typical data analyst, this meant a time consuming process to configure an application to prior to starting the even more time consuming process of analyzing data. The design, though new, felt dated and was too visually dense to work on a smaller screen I decided to take a radical approach toward the redesign.
Oracle - Big Data Discovery My proof of concept eliminated all chrome from the applications, gave it a new color palette to build personality, but most importantly, broke down many of the layout conventions that Endeca had used for years, and built common components into the framework of the application
When I presented this design to the leadership, the reaction was beyond anything I could have imagined. Despite being about to release the new redesign, we were mandated to collaborate as a department on a branded, modern design which incorporated the changes to the framework that I had proposed.
Oracle - Big Data Discovery
Examples of Visual Treatments. The winner is on the bottom right.
Oracle - Big Data Discovery Once the winning design was approved, the user exerpience for all of the pre-existing components had to be redesigned. We made a lot of wireframes.
Oracle - Big Data Discovery
But just before the finished product was to be released, we got word that we would not be releasing a new version of Endeca Information discovery........
.........Because, the design was too good and they didn’t want to waste it.
Oracle - Big Data Discovery
The Solution
The leadership had proposed to use Endeca’s technology to create a new product that would be billed as a data discovery application targeted at data scientists using Hadoop clusters. There were a few companies starting to address this market, but each only tackled a portion of the problem space. To effectively create a holistic solution, we had to create three new application areas: Catalog, Explore, and Transform. The previous application lived on as Discover.
Oracle - Big Data Discovery
Who are data scientists? Imagine a person with a Ph.D in Computer Science, and an MBA, but whose passionate is statistics. These are highly specialized, highly skilled people who are being hired by many large organizations to make sense of the ever-increasing amount of data that businesses collect. They use a varied toolset: Excel, R, Pandas, Tableau and others. Their workflow is highly iterative and quite time consuming mostly because of the nature of the data they work with. Most data is dirty. It is badly formatted. It may have rows of missing values. It may have abbreviations, blank spaces after entries, misspellings, etc. One data set may have names of cities; another might only have ZIP codes. All of these data quality issues must be dealt with before analysis may begin. Data scientists spend 80% of their time cleaning data. When the data is cleaned, it still might not be useful, and the data scientist has no way of knowing in advance whether the data set is worth putting effort into. When dealing with huge data sets, this problem is compounded. A data set must be ingested, sampled, cleaned using an ETL tool, then exported and ingested into the analysis tool. If there are still problems, the user must run the same cycle again. In a data set with a billion rows, this may take several days just to process. This is why many visualization providers don’t gain traction with this market. By the time they are done with the process, they just want to see something. If they have to use yet another tool, it adds more pain to the process.
Oracle - Big Data Discovery
Analytical Lifecycle
Image by Joe Lamantia
Oracle - Big Data Discovery Oracle Big Data Discovery addresses this lifecycle through four application areas:
Catalog:
Hadoop is a fairly opaque file system. It has no real UI. The Catalog solves this by providing a visual face to Hadoop. Inside the catalog are tiles for all of the data sets in the Hadoop Cluster, and all of the Oracle Big Data Discovery projects which access them. Whereas many Hadoop users interact with their system through a command line, the Catalog gives the users the ability to refine the data sets in view using Endeca’s search technology.
Explore
Explore is a lightweight visualization tool with which a user can examine the shape of their data and begin to understand the story that the data tells. Users can examine the condition of the data, for nulls and type mismatches, and begin to create visualizations by combining attributes into a scratchpad. The user gains efficiency by knowing that they are working with good, and relevant data before they start to analyze.
Transform
The Transform feature allows the user to clean the data without leaving their analysis environment. It also allows the user to enrich their data, and also to create new attributes. The sum of all of these transformations is contained in a script which may be applied to a sample of the data quickly, or to the entire data set without inturrupting their workflow. I was responsible for this section of the application.
Discover
The original Endeca product enhanced with addiional visualization types. The user confiures visualizations on pages, and uses guided navigation to drill into the data.
Simplified Site Map
Oracle - Big Data Discovery Again, we made a lot of wireframes. We worked closely with development to refine our ideas and overcome tehnical challenges. (Catalog wires by Lena Shukel, Explorer wires by Diana Ye)
Oracle - Big Data Discovery We then made further adjustments to the framework and visual designs.
Catalog
(Design by Lena Shukel)
Oracle - Big Data Discovery
Catalog - Data Set View Showing Additional Meta Data
(Design by Lena Shukel)
Oracle - Big Data Discovery
Explore
(Design by Diana Ye)
Oracle - Big Data Discovery
Transform - Record View
Oracle - Big Data Discovery
Transform - Faceted View
Oracle - Big Data Discovery
Transform - Transformation Editor
Oracle - Big Data Discovery
Transform - Transformation Script Panel
Oracle - Big Data Discovery
Discover - Showing Parallel Coordinates Chart, Available Refinments and Add Component Panel
Oracle - Big Data Discovery
The Result After about 10 months of design and development effort, the product was previewed at Oracle Open World in October 2014 to an enthusiastic reception. We were featured in several sessions as well as showcased in Thomas Kurian’s keynote. Endecca’s technology finally got the reception it deserved.
Chevron - Energy Trading and Information Management
Chevron -
Energy Trading and Information Management Concept Design Lead, UX Director
Chevron - Energy Trading and Information Management
Background Everyone knows Chevron as an oil company, but a large part of what they do is actually trading energy. They operate a huge trading floor and manage several systems which provide market information to their traders. Having access to the best and latest information is critical, and a loss of a system can be catastrophic. Given their importance, it was a surprise to us at Sapient that the health and status of these systems were being managed with an Excel spreadsheet. And not even a good one. Surely there was a better way, and I lead the charge to find it.
Chevron - Energy Trading and Information Management
The Problem
There were two separate issues we wanted to tackle: 1. The traders were viewing the availability of their data using a big screen TV with a presentation that took up a lot of space, but didn’t provide a lot of information that was relevant to them.
Chevron - Energy Trading and Information Management 2. The health and status of the systems which provided the information was managed in a very complicated Excel spreadsheet which was neither scalable, nor readable. There was no obvious connection between the systems, the business processes they represented, and the information the traders were looking at. Most of the spreadsheet was devoted to showing the completion times of various processes, but it wasn’t clear why, and there were so many data points that it was nearly impossible to see what data was relevant.
The Process To come up with a solution, we needed to talk with some subject matter experts, and luckily for us, the people managing these systems worked for Sapient as well. But they were in India, so we had a series of late night and early morning conference calls to get the information we needed.
In between calls, we sketched, first on paper, then onto white boards, and then in Illustrator working collaboratively to iterate quickly as new information became available to us. As our questions were answered, we gained a better understanding of the different user groups, and what each needed from our application.
The Solution There were three user groups for this system: 1. The traders on the floor, who needed an easy to understand representation of what data was current, what data was coming soon, and what data was going to be delayed. 2. The IT staff, who wanted to know what systems were running and which issues need to be addressed immediately. 3. Management, who needed to know the trending information behind which processes complete on time, and which ones tend to be late or break down, and why these things happened.
This required two different interfaces, a tablet/desktop app designed allow the user to explore data, and a read-only interface for the traders designed to be readable from a distance on a television. For the first interface, the challenge was to make explicit the connections between the systems and the business processes. As it turned out, each process was fed by several systems, and the health of that process was the aggregate of the health of the contributing systems.
The complicated trend charts that were unreadable in Excel, were meant to allow the user to see when processes were late, so we made an interface which allowed the user to set a tolerance for what can be seen, so only the relevant information is displayed. From there the user can delve into a particular incident when a system broke down, see what happened and why, and if necessary, take action to correct the issue.
The TV display for the traders was a little trickier, because we didn’t know which parameters were most important. So we made several versions, where we played with the way that the element of time was incorporated. Did the users just care if the data was ready or late? Was lateness something they wanted to track? Was the “nearness” of the data as relevant to the user as whether it was available now or not? These were questions that our contacts couldn’t answer, so we made many versions, and presented them as possible solutions to illustrate the importance of talking to the users.
The Result We packaged up all of our drawings, sketches, and thoughts into a presentation and sent it to our practice lead in India to present. What we wanted to show was our process, and approach, and not just the solution. As a result of this pitch, we were awarded a project to track the location of every barrel of oil in Chevron’s fleet as well as follow-up work to the ideas we had already pitched.
Fidelity Money Market Trading
Fidelity -
Money Market Trading Platform Information Architect, Visual Designer
Fidelity Money Market Trading
Background Fidelity trades a lot of money market securities, but the systems they use were archaic, complicated, and prone to performance issues. The interfaces they used ranged from Excel to tickets designed in Visual Basic, to Green Screen terminals dating back to the 1980’s. In 2012, Sapient was brought in to build them a new system.
The Problem: Different trading desks traded different securities which had different requirements for their trading tickets. Therefore we had to create a model that could be scaled to all ticket types.
Fidelity Money Market Trading
The Solution First we audited the existing UI’s to figure out what elements were common to all tickets. Based on our findings, we created several models of tickets, printed them out as posters and displayed them in a large conference room.
Fidelity Money Market Trading We then had the traders come in as a group and created an environment where they could walk from one poster to another, saying what they liked and didn’t like about each. Workshops were held with the traders at every stage of the project. This insured buy-in from all stakeholders throughout.
Fidelity Money Market Trading
When it was clear that there was a preferred model of ticket, we designed a more formal wireframe and presented it to the traders. Getting feedback along the way, we added additional value by including supplementary information in the tickets, so that they didn’t have to leave the experience to have the information they needed.
Fidelity Money Market Trading A formal deck was developed detailing every interaction as a guide for development. The project was developed using the Scrum methodology.
Fidelity Money Market Trading
Before long, it was clear that we could create auxiliary applications for situations where traders needed to launch a ticket quickly, and allocate later. These small applications were added to the scope of the project.
Fidelity Money Market Trading Fidelity requested that the system be available in user selectable themes. Visual designs were created for 2 themes. I then built and XAML prototype to demonstrate the interaction design
Fidelity Money Market Trading
In the end we created an entire suite of applications including tickets, a launcher application, a market surveillance monitor and a blotter, complete with visual designs in user selectable light and dark themes.
Fidelity Money Market Trading
The Result The system was adopted with the first tickets appearing in the summer of 2012. By the summer of 2013, over $750 billion of securities will be traded every day using the system we designed. The success of the project launched follow-up work with Fidelity for everyone involved.
Future Bank
Future Bank
The Future of Online Banking For Tablet UX Lead, Creative Director
Future Bank
Background At Sapient, we were often tasked with non-delivery projects to help stretch ourselves creatively. Oftentimes consulting is a matter of doing the best you can given the situation you are in. These projects gave us the chance to simply do the best we could.
Future Bank
The Problem: Right now you have no idea how much money you have. You don’t know what’s in flight, what is up or down, or your total net worth. For something that is so important to everyone, it would seem like that would and should be a concern. There are ways you can find out, but its complicated. But it doesn’t have to be.
Future Bank
The Solution: The concept of Future Bank is a platform agnostic, vendor agnostic web application which allows the user to see the behind the log-in information for any account they own. This is nothing new, applications such as FinanceWorks or Mint.com do much the same thing. But Future Bank takes it a step further by adding transactional functionality, giving the users the ability to act upon their entire financial picture even between different institutions.
Vitamin Water Blortal
Vitamin Water
The making of a “Blortal� Information Architect
Vitamin Water Blortal
Background Coca Cola wanted to promote Vitamin Water as a lifestyle brand, but didn’t know how.
The Problem: Vitamin Water is a soft drink, but Coca Cola didn’t want it to be associated with “soda”. They wanted it to be seen as a lifestyle choice for the youth of today, and associate it with music, sports, and fashion
The Solution: Many brands have a blog as part of their marketing, but its usually a side note to a portal of some kind. Why do you need both? By creating a blog-portal, we associated the brand with videos highlighting fashion, sports, and music, to make the association implicit without creating the feel of “advertising”.
Vitamin Water Blortal Speed was of the essence. The entire wireframe deck was produced over a 4 day sprint at Sapient’s Miami office.
Vitamin Water Blortal
The Result Coca Cola was psyched about their blortal. They hired Sapient for follow on work and a new version of the blortal was launched in 2013.
Miscellaneous
Miscellaneous Work
Miscellaneous
UBS Future trader Desktop Concept
Miscellaneous
Vistaprint Multi variant Product Customization
Miscellaneous
Fidelity Cross-Session Trading
Miscellaneous
Ipad Market Impact Concept