Research · Public transit
DC Student Travel
Privacy-preserving inference of metropolitan student mobility under partial observation.
Overview
This project reconstructs Washington, DC student travel without requiring student-level home address records. It estimates home-school demand, transit use, accessibility, and service impacts by combining multiple partial evidence sources.
The central question is practical and methodological: how much can be learned about student mobility when direct observation is incomplete and privacy must remain a hard constraint?
Method
- Generate a privacy-preserving synthetic student population with PASS.
- Infer event-related ridership from automatic passenger counts.
- Model plausible home-school travel with GTFS-based routing.
- Compare demand, observed transit evidence, accessibility, and service scenarios.
Research focus
The work distinguishes modeled travel from observational evidence, keeps uncertainty visible, and treats privacy-preserving reconstruction as a basis for evaluating transit access rather than as a substitute for directly observed individual behavior.