

RUJIE CHENG PORTFOLIO

EDUCATION
PERSONAL INFORMATION 2019 - 2024
Rujie Cheng
437 N 40th St Philadelphia, PA 19104
rujiech@design.upenn.edu
+1 (267) 648 - 9224
2024 - 2025
Junior Editor
Penn Panorama Journal
Philadelphia, PA
2021 - 2024
Graphic Designer
Institute of Humanities and Social Sciences, Peking University
Beijing, China
2020 - 2023
Project Manager
Peking University Public Space Design Project Beijing, China
SKILLS & LANGUAGES
Master of City Planning
Concentration: Housing, Community & Economic Development
University of Pennsylvania
Philadelphia, PA
Japanese (Intermediate) 2024 - 2026
Bachelor of Laws in Sociology
Peking University Beijing, China
ArcGIS Pro | Adobe Photoshop | Adobe Illustrator | Adobe InDesign | R | Python | STATA | SPSS | Microsoft Suite
Mandarin Chinese (Native)
English (Fluent)
4 Study of COVID Trends and Resource Deployment
Redistricting Pennsylvania’s Congressional Boundaries
12 Planning IndeGo Bike Share Stations
16 Brewerytown Neighborhood Profile (Excerpt)
20 Other Works
Study of COVID Trends and Resource Deployment
Project type: Individual Course Work
Date: December, 2024
Instructor: Shea O’Neill
Part 1 | Identifying COVID Clustering Trends
The maps below illustrate the clusters and outliers of New York City ZIP Code Tabulation Areas (ZCTAs) for COVID case rate, death rate, and the proportion of the population testing positive. Both confirmed and probable cases/deaths are included in the case/death rates. I applied the Contiguity Edges/Corners method to identify these patterns.
Overview
This project studied COVID-19 trends in New York City and optimized resource allocation in The Bronx. It first identified clustering trends for COVID case, death, and positive rates. It then examined three scenarios for Pandemic Response Center allocation, balancing access, distance, and case density. Finally, it conducted a detailed analysis of demographic, zoning, facilities & transportation, and parcel trends within a 1-mile walkshed of each proposed site generated under the second scenario.
Overall, the clustering patterns for the COVID case rate and percentage positive show similar trends: Midtown and Lower Manhattan exhibit low-low clusters, while Staten Island displays high-high clusters. However, despite the
Clusters/Outliers for COVID Case Rate
Tools Used
R | ArcGIS Pro |
Adobe Illustrator |
Adobe InDesign
Data Source
ACS 2023 5-Year Estimates
NYC Open Data
TIGER/Line Shapefiles
Clusters/Outliers for COVID Death Rate

high COVID case rate clustering in Staten Island, a similar pattern does not emerge for the death rate. Instead, high death rates are clustered in south Brooklyn.
In terms of Outliers, the COVID case rate shows a high-low outlier in Midtown Manhattan (10036) with a rate of 13,351 cases per 100,000 population. For the death rate, a high-low outlier appears in Lower East Manhattan (10002) with 463 deaths per 100,000 population.
The discrepancies between cluster and outlier patterns for COVID case/death rate can be partially attributed to age distribution. As shown on the map on the right, the highlighted areas with blue boundaries—either high-high death rate clusters or high-low death rate outliers despite low or insignificant case rates—generally have over 20% of their population aged 65 or older. This age group is more vulnerable to severe outcomes from COVID compared to younger populations, making these areas stand out from their neighbors.
Clusters/Outliers for % Positive

Percentage of Population Aged Over 65 by Census Tracts



Part 2 | Mapping Scenarios for Pandemic Response Center Allocation
This section focuses on just The Bronx and analyzes 3 scenarios for Pandemic Response Center allocation. Each scenario is defined by specific constraints and aims to optimize service outcomes within those constraints. The maps below illustrate the allocation.
Scenario 1
30-Minute Walk Accessibility for All Residents
Scenario 2
3 Sites / Minimal Distance for All Residents
Scenario 3
3 Sites / Minimal Distance for Susceptible Groups

The first scenario requires a total of 19 sites. The map shows Pandemic Response Centers over census tracts colored by their total population. Each red dot represents 1,000 residents. Generally, the site locations correspond closely to the population distribution. Sites are more likely to be situated within or near census tracts with populations exceeding 6,000 and are particularly concentrated in the densely populated southwestern Bronx.
For the second scenario, I applied the Minimize Weighted Impedance method. The three selected sites are located in the upper north, southwest, and middle Bronx, serving residents in the northern, southwestern, and southeastern parts of the Borough, respectively.
In the third Scenario, site allocation is weighted by the number of COVID cases. I also applied the Minimize Weighted Impedance method.
When prioritizing COVID cases instead of total population, the proposed sites generally moved westward. Particularly, the upper north site moved closer to northwestern Bronx, which had higher COVID case counts. Although northeastern Bronx has a larger population, its relatively lower number of COVID cases justified relocating the site to better serve the more susceptible population.
1-Mile Walkshed Around Proposed Sites in Scenario Two

Part 3 | Analysis of Trends around Proposed Sites in Scenario Two
(Excerpt)
This section presents an excerpt from the original analysis, providing an overview of demographic, zoning, facilities & transportation, and parcel data within a 1-mile walkshed of the three sites generated under Scenario Two. It highlights common trends observed across the 3 sites and includes an analysis of the 1-mile walkshed surrounding Site A. The analyses of Sites B and C are omitted, as their structures and layouts are similar.
• All three areas have large populations of around or over 100,000, with diverse racial composition predominantly made up of Hispanic/Latino and Non-Hispanic/Latino Black residents. This diversity highlights the importance of adopting multilingual resources at the Pandemic Response Centers. Site B stands out with an exceptionally high poverty rate of nearly 34%, significantly surpassing Sites A and C.
• The areas are primarily residential with public transit access. However, while Sites A and B benefit from adequate healthcare facilities, Site C shows a scarcity of healthcare resources, posing a potential challenge.
• Across all sites, there is a notable disparity between parcels’ market values and their actual assessed values (as estimated by the NYC Department of Finance), raising concerns about potential housing affordability issues.
Site A is situated in the upper north Bronx, near the eastern edge of Bronx River Forest. Its proximity to the Gun Hill Rd Subway station provides convenient access via public transit. The area surrounding the site features a diverse population composition, economic disadvantage, and adequate healthcare facilities. Parcel data highlights economic disparity, with high market values but low assessed values.
• The diverse racial composition suggests the Response Center prioritize bilingual resources to effectively engage communities.
• The area is mainly residential with several scattered parks. Two commercial corridors traverse the area. Manufacturing zones occupy a small section in the northwest.
• The area is well-served by healthcare facilities, enhancing the center’s capacity to respond quickly to patients and potential infections. A police station is located on the southwest periphery, but no public safety services are available within residential areas.
• While the high market value indicates real estate potential, it may also increase financial pressures on residents.

Redistricting Pennsylvania’s Congressional Boundaries
Project type: Individual Course Work
Date: December, 2024
Instructor: Shea O’Neill
Part 1 | Measuring Racial Diversity
Overview
In 2018, the Pennsylvania Supreme Court overturned the state’s congressional boundaries (drawn in 2011) to address gerrymandering. This project applies GIS to compare the old and new district boundaries based on racial diversity and compactness within each district. It also includes an effort to redraw three congressional districts in Southeastern Pennsylvania, aiming to balance population size, racial diversity, and the compactness of district shapes.
To measure racial diversity within each Congressional District, I applied Simpson’s Diversity Index1, which calculates the probability that two randomly selected individuals belong to the same racial (or ethnic) group. The value of Simpson’s D ranges from 0 to 1, where 0 represents infinite diversity and 1 represents no diversity. For this project, the population is categorized into the following groups: Hispanic/Latino, non-Hispanic/Latino White, non-Hispanic/Latino Black, nonHispanic/Latino Asian, and all other non-Hispanic/Latino races.
In Pennsylvania, cities and their surrounding suburban areas (e.g., Pittsburgh, Harrisburg-Lancaster-York region, and Philadelphia) generally exhibit higher levels of diversity compared to rural areas
Racial Diversity Distribution of Pennsylvania by Census Tracts
Tools Used
R | ArcGIS Pro |
Adobe Illustrator |
Adobe InDesign
Data Source
ACS 2023 5-Year Estimates
Pennsylvania Spatial Data Access

1 Simpson, E. H. (1949). “Measurement of Diversity”. Nature. 163 (4148): 688.
Racial Diversity and Total Population by Old Congressional Boundaries

The two maps on the left visualize racial diversity by Congressional District. Lighter colors indicate a higher degree of racial diversity. Since the geographic area of each district is not proportional to its population, I use graduated symbols instead of coloring the entire area, as coloring large but sparsely populated areas could mislead viewers into overstating the extent of low diversity across the state.
For the table below, I assigned letters to the old and new Districts based on their geographic similarity and proximity and used matching color to visually represent these groupings for clarity.
Overall, the Simpson’s D values for the Congressional Districts fall within a middle range, meaning the probability of two randomly selected individuals belonging to the same group is between 0.40 and 0.80. The southeast part of the state (i.e., Philadelphia and its surrounding suburban areas) exhibits higher levels of diversity, while the mid-western regions show the lowest diversity.
Racial Diversity and Total Population by New Congressional Boundaries
Comparing Simpson’s D values for the old and new boundaries, 10 out of 18 Congressional Districts exhibit relatively stable diversity levels (changes of less than ±0.03); 4 districts have experienced a significant decrease in diversity, while another 4 show a more heterogeneous racial composition.
Simpson’s D Value for Each Congressional District




Part 2 | Measuring Compactness
Compactness is another crucial criterion for Congressional Districts, as it reflects whether gerrymandering may have influenced the determination of district boundaries. A polygon’s compactness can be measured using the Polsby-Popper (PP) Score1, calculated by multiplying the polygon’s area by 4π and dividing it by the perimeter squared. The PP Score ranges from 0 to 1, with values closer to 1 indicating a shape that more closely resembles a circle, and thus greater compactness.
The table on the top right presents the PP Scores for each Congressional District under the old and new boundaries. 16 out of 18 districts have improved scores, meaning that their shapes are less irregular under the new boundaries. The average PP Score for the new districts is 0.33, compared to just 0.17 for the old districts. The median score for the old districts is even lower at 0.14, highlighting that most districts had very low compactness, with only a few high outliers skewing the average.
The 2018 redistricting altered several districts’ boundaries to improve compactness. The two maps on the bottom left focus on District C. While the old and new boundaries for this district share a substantial overlap, the PP Score for the new boundaries increased to 0.28, compared to just 0.15 previously. The score indicates that the district’s shape now occupies 28% of the area of a circle with the same perimeter, reflecting a notable enhancement in compactness.
Examining the inclusion and exclusion of counties under the old and new districting plans also highlights issues of gerrymandering. As shown in the map on the bottom right, the old boundaries bypassed Wyoming County but included several counties in Northeast Pennsylvania, creating a wide “notch” in the district’s shape. Additionally, only parts of Tioga, Lackawanna, Monroe, Northumberland, and Perry Counties were included, contributing to the old district’s irregular shapes.
In contrast, the new boundaries of District C aligns more closely with county lines, except for dividing Northumberland County in the middle. This approach results in a more coherent and justifiable rationale for the districts’ design.
PP Score Value for Each Congressional District
1 Polsby, D. D.; Popper, R. D. (1991). “The Third Criterion: Compactness as a Procedural Safeguard Against Partisan Gerrymandering”. Yale Law & Policy Review. 9 (2): 301–353.


Case Analysis: Change in the PP Score of District C

Part 3 | Redistricting Plan
The map on the right illustrates a proposed redistricting plan with 3 new Congressional District boundaries drawn in eastern Pennsylvania near Philadelphia. With the state’s total population reaching 13,000,000 in 2020, each Congressional District should ideally have around 720,000 residents. Using this guideline and existing county boundaries, I delineated three districts encompassing Bucks County and Philadelphia County. The table below summarizes their key properties.
Properties of the Proposed Districts
District X includes all of Bucks County and part of the northeastern Philadelphia, with a population of 740,086 and a Simpson’s D value of 0.64. Bucks County exhibits a relatively low level of racial and ethnic diversity, but the addition of northeastern Philadelphia increases both the total population and the district’s diversity.
District Y spans the northern part of Philadelphia. It has a population of 768,877 and a Simpson’s D value of 0.29, indicating a densely populated and highly racially diverse area.
District Z covers the remainder of Philadelphia, with a population of 719,462 and a Simpson’s D value of 0.32, also reflecting racial diversity.
To conclude, the proposed districts maintain populations slightly higher than those in central and western Pennsylvania. The boundaries primarily align with county lines and achieve fair compactness, as indicated by their PP Scores. This ensures both population balance and geographic cohesiveness, while also reflecting the demographic diversity of the state.

Planning IndeGo Bike Share Stations
Project type: Individual Course Work
Date: October, 2024
Instructor: Shea O’Neill
Part 1 | Analyzing Current IndeGo Usage
Overview
This project analyzes IndeGo trip data of Q2, 2024, along with U.S. Census data. The first part of the report summarizes and visualizes IndeGo bike share trends across Philadelphia. The second part provides recommendations for future IndeGo station placements, based on an analysis of local demographic, socio-economic, and spatial characteristics.
Launched in 2015 by the City of Philadelphia, IndeGo now has over 250 stations across the city, providing sustainable and convenient transportation options for both visitors and local residents. While initially concentrated in the city center, the network has gradually expanded into less central areas over the past decade. However, as of 2024, Northeastern Philadelphia remains without active IndeGo stations. The table on the right presents the most popular stations (with highest number of trips that start/end at each station) by year of activation.
Tools Used
R | ArcGIS Pro |
Adobe Illustrator |
Adobe InDesign
Data Source
ACS 2022 5-Year Estimates
IndeGo Trip Data
OpenDataPhilly

Most Popular IndeGo Stations in Q2, 2024 by Year of Activation
IndeGo Stations by Year of Activation
2023 17th & Locust 8440
2024 13th & Locust 6824
• Rittenhouse is the most popular neighborhood with a total of 92,641 trips.
• When calculating trips per station, Rittenhouse still ranks first, with an average of 6,176 trips per station.
• Major starting stations are clustered in Center City and University City
• Most IndeGo stations have relatively equal number of outgoing and incoming trips
• Longer trips usually start from the northern part of Schuylkill River banks and West Philadelphia (measured by median commute time).



Part 2 | Planning Future Stations
To recommend locations for future IndeGo bike station implementation, three key criteria were analyzed at the census tract level:
• Biking Preference
Census tracts with a relatively higher proportion of commuters using bikes compared to other areas in Philadelphia were prioritized. A threshold of 6% was established based on citywide trends, and tracts exceeding this threshold were considered ideal candidates for future bike station placement.
• Current Shortage of Bike Stations
Census tracts that currently lack IndeGo bike stations within their boundaries were identified as areas of unmet need..
• Proximity to Parks
Census tracts with parks either within or near their boundaries were highlighted. Proximity to green spaces supports cycling for recreational purposes, expanding bike usage beyond commuting to include exercise and leisure.
Based on these criteria, three census tracts were selected as candidates for future bike station installations, as shown on the map. A detailed analysis of each tract and specific recommendations for future bike station locations are provided on the next page.


The first zone is located in Southwest Philadelphia, at the northeast corner of the Paschall neighborhood. While there are few IndeGo stations nearby, 7.42% of all commuters in this zone use bikes, indicating significant demand for bike-sharing initiatives. With three parks in close proximity, this zone is ideal for future IndeGo stations, particularly along its border near the parks.
The second zone is located in Point Breeze, a multi-cultural neighborhood in South Philadelphia. This zone has an exceptionally high proportion of bike commuters, which exceeds 20%. Although two IndeGo stations are located near the zone’s periphery, adding new stations in the center of the zone would better support local commuters. Placing future stations near the park within the zone would be a good fit for the area.
The third zone is situated along the Schuylkill River, spanning Fitler Square and southwestern Rittenhouse. Here, 13.7% of commuters use bikes. Its prime location next to Markward Playground and close to the city center makes it ideal for new IndeGo stations, which would serve not only bike commuters but also tourists and exercisers.


Brewerytown Neighborhood Profile (Excerpt)
Project type: Group Course Work
Date: September - December, 2024
Instructor: Xiaoxia Dong
Overview
This 4-person group project aimed to comprehensively analyze and narrate the story of Brewerytown within the context of Philadelphia, synthesizing its current conditions, changes since 2012, and future projections for 2030. My contributions to this project included writing the sections on education, employment, and income & poverty, as well as editing graphics and formatting the layout for the entire report book.
This report provides a comprehensive analysis of Brewerytown, a neighborhood located in the Lower North District of Philadelphia. It examines the characteristics and transformations observed over the past decade (2012-2022) across key areas such as history, demography, education, economic development, household composition, rent, income and poverty, employment, and commuting patterns. The report also identifies the opportunities and challenges currently facing the neighborhood.
The findings highlight an ongoing process of gentrification that is reshaping Brewerytown. The long standing low-income Black community that has been living in the neighborhood for the past 50 years has experienced pressure in the past decade due to the surge in development, influx of younger residents, and increasing rental costs. This threatens to uproot the core community of the neighborhood.
Tools Used
R | ArcGIS Pro | Microsoft Excel
Adobe Illustrator | Adobe InDesign
Data Source
ACS 2022 5-Year Estimates
OnTheMap
OpenDataPhilly



Education
One of the greatest changes in neighborhood outcomes between 2012 and 2022 for Brewerytown is education. Residents of Brewerytown are now much more likely to have a bachelor’s or master’s degree. Among residents aged over 25, the share of those holding a bachelor’s degree saw an approximate ten-percentage-point increase from 15.1% to 26.5%, and the residents with a master’s degree or higher increased from 17.2% to 26.7%. The growth in overall educational attainment level in Brewerytown outpaced the citywide trend, as the bachelor’s share grew from 13.7% to 18.8%, and graduate degree holders from 9.6% to 14.8% in Philadelphia (Figure 1). The level of higher-level education in Brewerytown started off with higher rates than Philadelphia (in 2012) and has continued to grow at faster rates than Philadelphia. Geographically, Brewerytown sits between Temple and University City where both Drexel and the University of Pennsylvania are located. The 49 bus route connects Brewerytown to University City, which was introduced in 2019. Thus, populations from both areas who either work or attend the Universities live in the neighborhood and could contribute to the increasing levels of higher educational attainment.

Additionally, the number of residents with only lower educational attainment dropped. Those with a high school education with no diploma decreased from 13.5% to 6.9%, and the number of residents with only a high school diploma also dropped from 31.2% to 20.6%. In sum, less than 10% of residents in 2022 had no degree at all (high school, college, etc.), compared to 16.8% of residents in 2012.
Employment
Among Brewerytown’s nearly 14,000 residents over 16, the labor force participation rate is 74.7%. Over the past decade, the neighborhood’s labor force has expanded by 45.8%, with the 25-to-44 age group seeing the largest growth, increasing by 83.8%. Both the unemployment rate and the proportion of residents over 16 who are not in the labor force have dropped by more than 10 percentage points, far exceeding the citywide trend (Figure 2). The inclusion of non-labor force participation acknowledges the factor of “discouraged workers” in the economic vitality (or lack thereof) of an area. In sum, Brewerytown is contributing more to the Philadelphia economy than in the past few decades. The influx of working-age residents indicates a significant demographic shift and economic revitalization taking place in the neighborhood.
Figure
Unemployment Rate & Population Over 16 Not in Labor Force

However, while the unemployment rate among Black residents has dropped by 57.8%, it remains above 10% (11.5%), which is 4.5 times that of White residents (2.5%). This indicates the uneven employment improvements across different racial groups.
A high proportion of Brewerytown’s workforce is employed in professional or white-collar jobs. The largest occupational group in the neighborhood is people working in management, business, science, and arts occupations, with 59.4% of residents in the labor force employed in these fields. Additionally, 14.8% of the working population is employed in sales and office roles. This further underscores the fact that Brewerytown has a predominantly young demographic who are well-educated, working professionals.
On a per capita basis, the average income for White residents is times that of Black residents. 2.2
The median income of households with a White householder is times the amount of households with a Black householder. 2.7
Income & Poverty
Over the decade from 2012 to 2022, Brewerytown experienced significant household median income growth that far outpaced the citywide trend. The neighborhood’s median household income increased by 71.1% from $42,835 in 2012 to $73,307 in 2022 (in 2022 inflationadjusted dollars). In comparison, the median household income in Philadelphia grew from $45,913 to $57,537 (25.3%). Moreover, both the absolute number and the percentage of residents living below 150% of the poverty level have declined in Brewerytown. Residents at or above this threshold have increased by 46.1% in number and 11.1 percentage points in share.
Despite the apparent prosperity in income amounts, disparities persist among residents. There are still 29.1% of Brewerytown’s households falling below the $34,999 income threshold; compared to Philadelphia, Brewerytown has a similar proportion of lower-income residents, but
a significantly larger share of wealthy residents (annual household income over $150,000) (Figure 3).
Families with children are facing financial hardships at a higher rate. Among families with children under 18, 38% live below the poverty level, and the poverty rate is even more concerning for families with children under 5, with 24% falling below the poverty line. In contrast, marriedcouple families are financially more stable, with only less than 2% of them struggling under the poverty line.
White households in Brewerytown are more concentrated in higher income groups. Whereas Black households show a more dispersed income distribution, with a notably larger share in lower-income brackets compared to White households. The median income of households with a Black householder is about $37,200, while households with a White householder have a median income of $100,700. However, this trend may not be as apparent when analyzed at the census tract level.

of Report Book Spreads
Although the selected census tracts span far less than a mile from end to end, there are distinct socio-economic differences between the tracts. For instance, the number of households with an income over $200,000 was 429 in tract 136.02 (on the southern edge of Brewerytown) and just 21 for tract 149 (on its northern edge). Thus, although overall economic outcomes have improved for an aggregate of neighborhood households, certain tracts have not experienced such an increase. Acknowledging this condition is critical when determining an overall neighborhood portrait and how neighborhood changes vary across smaller geographies.
Looking closer at the overall income growth trend, Brewerytown has a pattern that different income groups grew at different scales. Changes in high-earning yearly income outpaced that of Philadelphia. For high earners (identified in this report as earning $75,000 or more each year) Brewerytown generally demonstrated huge increases compared to Philadelphia. For instance, the most dramatic increase was for Black or African American households earning $150,000 to $200,000. From 2012 to

2022, there was a 1,475% increase in that group. There also was a similarly massive increase for Black households earning $125,000 to $150,000 of 690%. On the other hand, increases in the median income of White households trended more closely with Philadelphia as a whole, or were lower.
On a numeric basis, the number of high-income earners ($75,000 + per year) for both Black and White households increased dramatically. The number of high-earning White households increased from 1,036 to 2,301. For Black households, the increase was from 359 to 1,133. Thus, from 2012 to 2022 high-income earning households grew in both major groups, with the growth being even more substantial among Black households. If trends continue, the gap in the number of high-income earning households between the Black and White demographics will continue to narrow. The increased economic outcomes for Brewerytown and both Black and White households within its boundaries represent a growing strength from 2012.

Other Works
GRAPHIC DESIGN



PHOTOGRAPHY



