Optimizing Transit Access Equity in the DC Metro Area
Advisor: John Winchester
Internship Location: Villanova University
Onsite Mentor: Chenfeng Xiong
This research project aims to optimize subsidy incentivization strategies by applying Equity Credit Methodologies among transit service providers in selected regions, in this case, the Washington Metropolitan Area, in order to increase accessibility. It explores the design of a fair subsidy distribution system to improve transportation equity and social well-being. In the United States, unequal access to the transportation system exacerbates social inequality. At the same time, heavy reliance on cars has also exposed the inaccessibility of public transportation, bringing side effects such as severe air pollution, ultimately leading to climate change and congestion. To close the gap between accessible transportation options and residents needing these services, our research focuses on how to design a fair subsidy distribution system for each income group, thereby optimizing transportation equity improvements as well as revenue for transportation service companies. We aim to construct a comprehensive framework that considers factors such as demographics, economic status, geographical location, and transportation accessibility of target areas and further analyze the impact of the proposed subsidy allocation. Our procedures include designing optimization models, collecting data, applying the models, and performing sensitivity analyses. Ultimately, the goal is to incorporate carbon credit measures into the subsidy system by allocating individual emission credits to citizens using public transport. This will allow companies that need carbon credits to receive them from individuals, thereby paying for subsidies from transport companies or governments. This comprehensive approach will not only help policymakers with subsidy allocation but also reduce inequalities in transport access. It will also reduce emissions and provide a model that can be replicated in cities throughout the world.
Project Blog
Week 1: Jiashu Week 1
Hi, my name is Jiashu. Welcome to my first Senior Project blog post!
I’m excited to share the beginning of a research journey that builds on my previous work and continues with the support of a fantastic team at Villanova University. I’ve been a part of this amazing research group since last summer, and together, we have completed many impactful research.
For this project, I will be developing a comprehensive statistical model aimed at optimizing subsidy incentivization strategies among transit service providers in the Washington D.C. area. Our goal is to design a fair subsidy distribution system that not only improves transportation accessibility but also addresses broader social equity issues.
I am excited about my project and look forward to sharing it with you during my final presentation.
Jiashu
Week 2: Jiashu Week 2
Hello, I'm Jiashu. Welcome to my second Senior Project blog post!
This week, I am collecting data for my model. To calculate costs of travel effectively, I collected information on various origin-destination (OD) pairs in Washington, D.C.
I selected ten zip codes in and around Washington, D.C. (20707, 20708, 20705, 20783, 20742, 20740, 20782, 20781, 20737, and 20770) as our origin and destination zones. I used a fuel cost calculator to calculate the cost of automobile travel between the zones. I then used Google Maps to find corresponding transit cost data. This method allows us to make comparisons and analyze the relative strengths of different transportation types.
I’m excited to analyze these data further and integrate them into our overall models.
See you next week,
Jiashu
Week 3: Jiashu Week 3
Hi, I'm Jiashu. Welcome to my third Senior Project blog post!
I spent this week gathering more important data that are needed in order to add depth to our statistical model. I gathered a range of various data that are important for our analysis, and they include primarily the American Community Survey (ACS) and Washington Metropolitan Area Transit Authority (WMATA) data.
American Community Survey (ACS):
ACS, conducted by the U.S. Census Bureau, surveys about 3.5 million households annually to gather the social, economic, housing, and demographic realities across the U.S. That comprehensive information tells us factors that lead to transit usage from the community side.
WMATA Data:
The WMATA data provided data on transit fare levels, routes in the Washington Metropolitan area, and ridership information.
Merging these data sets with the travel cost data I collected in Week 2, we are now well-positioned to construct an integrated model that accounts for both economic and demographic considerations as well as practical transit cost realities. This multi-dimensional approach will enable us to develop a subsidy system that is not only fair but also responsive to the diverse needs of the community.
See you next week,
Jiashu
Week 4: Jiashu Week 4
Hi, I am Jiashu. Here's my fourth Senior Project blog post!
This past week, I merged and standardized all of the data we'd collected into a single Excel sheet, as below:
• Page 1: Travel Times & Costs
Contains auto and transit costs in addition to travel times for all ZIP-to-ZIP origin-destination pairs.
• Page 2: Usage & Demographics
Includes population, median hourly earnings, number of trips per day, and other big variables based on ACS data.
• Page 3: Utility & Volume Calculations
Uses part 1 of our model to estimate automobile and transit utilities, value of time (VOT), and forecasted OD volumes. Also gives convenient metrics such as trip rate and total travelers between each OD pair.
With everything in order, we are ready for a more detailed analysis and model building in weeks to come.
See you next week,
Jiashu
Week 5: Jiashu Week 5
Hi, I am Jiashu. Welcome to my fifth Senior Project blog post!
I have worked this week very closely with my mentor to validate and finalize the equation of our model. After we finalized the equations, I added more data into the Excel sheet to complete the dataset.
Using Page 1 data (costs and travel times) and Page 2 data (frequencies and demographics), we calculated the base accessibility of the current transit system and base Theil index to measure inequality of access before any subsidy is applied. These serve as important reference points to measure the efficiency of our model.
I then ran our nonlinear optimization model to determine the final, more equitable subsidy distribution across all origin-destination pairs and used the result to estimate the base subsidy for each origin–destination pair.
With this modeling step completed, we’re now moving into deeper evaluation and comparison of outcomes.
See you next week,
Jiashu
Week 6: Jiashu Week 6
Hi, I'm Jiashu. Welcome to my sixth Senior Project blog entry!
This week, I spent time finishing up all of the datasets with the equations that we've developed throughout the project. One of the most important tasks was calculating the adjusted utility values for the autos and transit modes after applying the subsidies calculated by our nonlinear optimization model. These adjusted utilities form the foundation of our model's behavioral predictions.
We then estimated several key outcome measures:
Updated Theil inequality index to assess changes in equity
Transit agency revenue estimates based on ridership and fare changes
System accessibility scores to assess improvements between ZIP-code pairs
We now have most of our final results at this stage. The next few steps will involve interpreting these results and graphing the data to clearly display our findings. I look forward to sharing the information that we have learned in the last few weeks.
See you next week,
Jiashu
Week 7: Jiashu Week 7
Hi, my name is Jiashu. Welcome to my seventh Senior Project blog post!
This week was all about data processing, interpretation, and initial visualizations. After finalizing the key metrics last week, I began analyzing the results in detail to understand the broader impact of the subsidy model.
The first major finding is that the inequality index (Theil index) of the transit system significantly decreased after introducing the subsidy, while system accessibility increased. These two changes indicate that our model had a meaningful, positive effect—making transit access more equitable and improving overall reachability across the ZIP-code pairs.
To better understand the scale of this change, I created a comparison using several subsidy levels:
Base case (no subsidy)
Partial subsidy thresholds
Full or optimistic subsidy scenario
For each of these, I compared both the inequality index and accessibility scores to see how different levels of investment affect outcomes. This helped identify where major gains occur and whether there are diminishing returns at higher subsidy levels.
Lastly, I began creating summary statistics and graphs to support our findings visually. These early visualizations include line graphs showing changes in accessibility and inequality under each scenario, and charts that summarize key trends.
Next week, I’ll continue building out more visualizations and start preparing my presentation materials.
See you next week,
Jiashu
Week 8: Jiashu Week 8
Hi, I'm Jiashu. Welcome to my eighth Senior Project blog post!
This week, I visualized our model's results using various subsidy thresholds to examine their impact on significant variables: inequality index, accessibility, and transit agency revenue.
The visualizations easily show a clear pattern:
As the level of subsidy increases, the inequality index decreases
Accessibility rises, which implies that individuals have better access to opportunities
And transit revenue rises, with greater ridership and fare box collection
These results show that all stakeholders benefit from the subsidy—transit agency, riders, and communities.
To further illustrate this, I created a 3D graph of all three variables together at different levels of subsidy. The graph shows a convergence point at the optimistic level of subsidy, where inequality is lowest, accessibility highest, and revenue highest. This graph provides a compelling summary of our findings, showing that best results can be obtained across the board with strategic subsidies.
Next week, I’ll continue refining these visuals and begin drafting key takeaways for my final presentation.
See you next week,
Jiashu
Week 9: Jiashu Week 9
Hi, my name is Jiashu. Welcome to my ninth Senior Project blog post!
This week I shifted focus from data analysis to writing the research paper. I drafted the overall structure and began filling in each section. Here’s the current outline:
Abstract – A concise summary of our objectives, methods, key findings (inequality reduction, accessibility gain, revenue impact), and policy implications.
Introduction – Background on transit equity challenges in Washington, D.C., project motivation, and research questions.
Literature Review – Overview of prior work on subsidy design, equity metrics (e.g. Theil index), and transit accessibility modeling.
Methodologies – Description of our statistical choice model, the nonlinear optimization framework, and calculation of utilities and Theil indices.
Data – Sources and processing steps: ACS demographics, WMATA fares/routes, OD travel costs and times, and calculated utility measures.
Results – Presentation of base vs. subsidized scenarios: changes in Theil index, accessibility scores, and transit revenues, with summary tables and visualizations.
Conclusion – Interpretation of findings, limitations, and policy recommendations for equitable subsidy implementation.
Over the next week, I’ll expand each section, integrate figures and tables, and refine the narrative to ensure clarity and cohesion. I’m excited to turn our analyses into a compelling written report.
See you next week,
Jiashu
Week 10: Jiashu Week 10
Hi, my name is Jiashu. Welcome to my final Senior Project blog post!
This week, I focused on finalizing and editing the research paper and began creating the presentation for my final defense. Key activities included:
Paper Refinement – Polished each section for clarity, tightened the narrative flow, and ensured all figures and tables were properly labeled and referenced. I also incorporated feedback from my mentor on the discussion and policy recommendation sections.
Proofreading & Formatting – Checked for consistency in terminology (e.g., “accessibility,” “Theil index”), corrected typographical errors, and formatted the document according to Villanova’s thesis guidelines.
Presentation Development – Drafted slides covering the project motivation, methodology, data sources, key results (inequality reduction, accessibility gains, revenue impacts), and policy implications. I sketched initial designs for graphs and maps to include.
Rehearsal Plan – Outlined a rehearsal schedule with my mentor to practice delivery, anticipate questions, and refine talking points for a clear, engaging defense.
last post!!!
Jiashu