Last Mile Connectivity
Data-driven transit optimization for the DMV area.
tech stack
AWSAthenaPostgreSQLPythonGeopandas
The Problem
Public transport ridership in the DC/Maryland/Virginia area suffers from the "last mile" problem—commuters abandon transit if the walk to their final destination is too far or unsafe.
What I Built
Developed a spatial analytics platform to identify connectivity gaps between WMATA rail stations and Capital Bikeshare hubs. The system highlights underserved regions that need new bike docks or shuttle routes.
Architecture & Approach
Built a geospatial data warehouse using AWS Athena and PostgreSQL/PostGIS. Used Geopandas to perform station-to-hub proximity analysis across 1M+ trip records and US Census Bureau demographic data.
Impact & Results
Identified 12 specific DMV neighborhoods where bike-share expansion would yield the highest ridership lift.
Correlated bikeshare availability with a 7% increase in WMATA station utility.
Presented findings to regional transit stakeholders to inform infrastructure planning.
Key Decisions & Tradeoffs
Opted for AWS Athena for the primary analysis to handle massive CSV datasets without the overhead of maintaining a constant RDS instance.