Subway Challenge

How can we better predict and prevent subway delays and reduce the burden on customers?

New York City Transit serves an annual subway ridership of over 1.7 billion, runs nonstop 365 days a year and spans 472 stations. With 5.6 million passengers and over 8,200 trips per weekday on routes with multiple merge points, numerous types of incidents occur each day resulting in over 2,500 delays per weekday. Each delay requires NYCT staff to make complex decisions and communicate with customers in real-time. 

The Transit Tech Lab seeks applications from companies to better predict subway incident impacts and serve customers. Solutions might include opportunities to analyze train arrival data, incident information and social media, leverage machine learning, develop predictive models, and improve real-time customer communication.


Where can I access historical subway arrival time data?
The archive of GTFS-RT (General Transit Feed Specification Realtime) for lettered trains except the L in August and September 2018 is available at:

Where can I access base subway schedules?
The current base subway schedule is available as GTFS from

Where can I access subway routing data?
This information can also be obtained from the schedule and line strip maps at

Where can I access historical subway incident and delay data?
Delays are exposed in the MTA's real-time arrival information. Published service alerts are archived at

Where can I access planned service changes, known as supplemental schedules?
Planned service changes are incorporated in the Supplemented GTFS feed, at



Fast Forward: The Plan to Modernize New York City Transit

MTA Subway Performance Metrics Dashboard 

Base Subway Schedules, Train Arrival Times (GTFS-RT), and more

Historical Train Arrival Time Data for October 2018