Faculty Projects
Building an AI-Powered Urban Mobility Analytics Platform: A Hands-On Undergraduate Research Experience in Chicago

Ying Chen headshot

Faculty

Ying Chen, CEE

Amount Requested

$24,000

Summary

Urban mobility is being transformed by artificial intelligence and large-scale data — yet undergraduate engineers rarely get hands-on experience with these tools at real-world scale. This project builds an open-source AI-powered urban mobility analytics platform for Chicago, integrating multiple large-scale data sources including passive mobility data, road networks, census demographics, and land use information. Four undergraduate McCormick students will work alongside faculty to design, build, and deploy machine learning models that predict destination choice, measure accessibility gaps across neighborhoods, and visualize how distance and urban form shape where Chicagoans travel. The platform will be publicly released on GitHub and presented at a transportation research conference, giving students a rare portfolio-level achievement that demonstrates end-to-end AI engineering on a real urban problem. Beyond the technical skills gained, students will engage directly with one of the most pressing challenges at the intersection of AI and sustainability — understanding and improving how cities connect people to the places they need to reach. Murphy Society funding provides the essential seed investment to launch this initiative, generate preliminary research results, and build the infrastructure that will support future undergraduate cohorts and competitive federal grant applications for years to come.

Planned Activities/Investments

This project unfolds across three structured phases during the 2026-2027 academic year, each with clear activities, student roles, and measurable deliverables.

Phase I - Data Integration and Infrastructure (Fall 2026)

Activities:

  1. Recruit and onboard four undergraduate students.
  2. Conduct weekly training sessions on urban data science fundamentals - spatial data, road network analysis, census demographics, and passive mobility data processing.
  3. Integrate multiple data sources (passive mobility data, OpenStreetMap road networks, American Community Survey, Chicago Data Portal, GTFS transit feeds, EPA Smart Location Database) into a unified spatial database.
  4. Build and document a reproducible data processing pipeline covering origin-destination trip cleaning, network distance computation, and demographic merging.
  5. Establish a cloud computing environment for model training and platform hosting.

Phase 2 - AI Modeling and Analysis (Winter 2027)

Activities:

  1. Develop and train a progression of destination choice models - from interpretable gravity models to conditional logit to machine learning models.
  2. Build a comprehensive mobility security index that rates neighborhood access to key destination categories.
  3. Analyze distance decay patterns across income groups, car ownership levels, and neighborhood types to identify equity gaps in urban accessibility.
  4. Document all model code and results in a reproducible research repository.

Phase 3 - Platform Development and Public Communication (Spring 2027)

Activities:

  1. Design and build an interative web-based dashboard using Python Dash or Streamlit with three core views: Mobility Pattern Explorer, AI Mobility Security Index, and Distance Decay Explorer.
  2. Conduct user testing with McCormick faculty and urban planning stakeholders.
  3. Submit platform and findings to a transportation or urban AI conference for student presentation.
  4. Prepare project summary report for Murphy Society donors.

Impact

This project creates meaningful impact across three groups:

  • Undergraduate Students (Direct and Primary Impact)

Four McCormick undergraduate students will receive an exceptional research experience that goes far beyond typical coursework. Students will gain end-to-end competency in AI and data engineering applied to a real urban problem — skills increasingly demanded by employers in technology, consulting, government, and graduate programs but rarely developed at the undergraduate level. Specifically, students will leave the project with:

Demonstrated experience building and deploying AI systems on real large-scale data:

  1. A portfolio project showcasing end-to-end data science and software engineering
  2. Direct exposure to faculty-led research and academic publication processes
  3. Professional mentorship from faculty with active research and industry connections

The project also establishes a reusable research infrastructure and mentorship template that benefits future undergraduate cohorts well beyond the funding period.

  • The McCormick School of Engineering

The project directly advances McCormick's strategic priorities in AI and sustainability by producing a publicly visible research artifact — an open-source platform — that demonstrates McCormick's capacity to train undergraduates on cutting-edge AI methods applied to societally important problems. Conference presentations and GitHub visibility raise the school's profile in the urban AI and transportation research communities. The project also strengthens the faculty member's external grant portfolio, supporting McCormick's research mission beyond the funding period.

  • The Chicago Urban Planning and Transportation Community

The platform produces actionable insights about neighborhood-level accessibility gaps in Chicago — identifying which communities face the greatest barriers to reaching essential destinations. These findings will be shared directly with the City of Chicago Department of Transportation (CDOT) and the Chicago Metropolitan Agency for Planning (CMAP), giving urban planners a data-driven tool to prioritize infrastructure investments, transit improvements, and land use decisions that reduce travel barriers for underserved communities. While the platform is not a policy instrument itself, it provides the evidence base that informs more equitable and sustainable planning decisions.

Deliverables

Phase 1: Clean, documented multi-source urban mobility dataset for Chicago, ready for AI modeling.

Phase 2: Validated AI models with documented performance metrics and preliminary research findings suitable for inclusion in a faculty journal paper

Phase 3: Publicly deployed open-source platform, conference presentation, symposium poster, and donor summary report.

Sustainability

The platform and preliminary results generated during 2026–2027 will directly support competitive grant applications to federal agencies.

Specifically:

National Science Foundation (NSF) — programs supporting AI, urban systems, and community-focused research remain active funding priorities at NSF, and preliminary results from this project will directly strengthen future faculty grant applications in these areas

USDOT University Transportation Centers (UTC) — destination choice modeling and network accessibility are core UTC research themes

Budget Overview

  • Undergraduate Student Stipends: $16,000 — Four undergraduate research assistants at $4,000 each for full academic year involvement across all three project phases.
  • Cloud Computing and Data Storage: $2,500 — Cloud infrastructure for platform hosting, model training, and storage of large-scale urban mobility datasets throughout the project period.
  • Data Access: $1,500 — Supplementary commercial mobility data license fees for research use, complementing freely available open data sources.
  • Travel and Dissemination: $3,500 — Support for students to present project findings at McCormick undergraduate research symposium and regional transportation or engineering events.
  • Supplies and Materials: $500 — Poster printing, presentation materials, and project documentation for symposium and donor reporting.

Total Budget Amount: $24,000

Matching Funds

At this time, no formal matching fund commitments have been secured for this project.