This project provided programming and analysis support for the creation of WMATA’s Bus Ridership Datamart and Bus Priority Datamart. WMATA has had unsupported access to fine-grained automatic vehicle location (AVL) data for years and needed help loading, cleaning, and aggregating this data so their planners could use it in bus priority decision-making. They also needed help cleaning raw automatic passenger count (APC) data.
Together with our partners, Foursquare ITP wrote a number of algorithms to match the fine-grained AVL data to the road network, parse it into travel time decomposition buckets, make vehicle heading data useful, and create fields for easy aggregation. We also helped create useful dashboards for staff to explore and download APC data. As a result, WMATA planners now have access to a rich new data source to conduct before/after analyses for transit priority treatments, select new locations for bus priority treatments, and gain insight on ridership from their APC data.
As a result, WMATA planners now have access to a rich new data source to conduct before/after analyses for transit priority treatments, select new locations for bus priority treatments, and gain insight on ridership from their APC data.
Foursquare ITP was part of a team that provided programming and analysis support for the creation of WMATA’s Bus Ridership Datamart and Bus Priority Datamart. Both datamarts consume raw APC and AVL data, including second-by-second fine-grained AVL data known as rawnav. Our team led the technical components of the Bus Priority Datamart, including several key rawnav data processing efforts in Python:
- Calculation of speed and runtime;
- Decomposition of bus travel time into different buckets, including things like bus stopped with doors open, bus accelerating, bus decelerating, and bus stopped but not serving passengers, among others;
- Translation of bus heading data into turns and calculation of angular speed;
- Map matching of data to the Open Street Map network; and
- Various other data cleaning processes.