Datasets
Every table we publish, with its size and what one row means. Rows and sizes come from catalog.json each time the docs are built.
Subway
Subway segment timesDaily run time between every pair of stations in a row, weekdays since April 2021.Subway slow zonesEvery stretch where trains ran well over their usual time for 4+ weekdays, since 2022, with the hours it cost.Slow zone hours by weekTrain-hours and rider-hours lost to slow zones across the subway, week by week.Feeds archive
Subway departuresOne row per train per station, written as the train leaves, from our own collector.Bus positionsEvery bus's reported position, about every 30 seconds, stored as change events.Bus arrival predictionsThe arrival times the MTA predicted for each bus at each stop: the feed behind countdown clocks.Service alertsMTA bus, subway and rail service alerts, one row each time an alert appears, changes or ends.Elevator and escalator outagesEvery MTA elevator and escalator outage, one row each time an outage appears, changes or ends.Elevators and escalatorsThe MTA's list of elevators and escalators, kept as change events.Bus
Bus route speedsOur bus speeds by route, direction and hour, one set of rows per day, from bus positions.Bus delay and bus lanesEach stretch between two timepoint stops: how slow it is, the time buses lose, and how much has a bus lane.Bus delay by streetBus time lost added up by street, across routes and directions, with bus lane coverage.Trips
Yellow taxi tripsEvery yellow taxi trip since January 2023, cleaned, with zone names and readable codes.Green taxi tripsEvery green cab (street hail livery) trip since January 2023, cleaned, with zone names.Uber, Lyft and Via tripsEvery trip by the high-volume for-hire services since January 2023, cleaned, with zone names.Two kinds of table
Section titled “Two kinds of table”Ours. Subway, bus and feed archive tables are built by us from the MTA’s feeds and published data. They are CC BY 4.0.
Trips. Taxi and Uber/Lyft trips are the TLC’s public files, cleaned and with zone names added, kept here so you can join them with ours. They stay under the city’s terms. We don’t carry Citi Bike trips; here’s why.
Where the files are
Section titled “Where the files are”| Layout | Path |
|---|---|
| Small tables, one file | https://data.transitlab.nyc/v1/<table>.parquet |
| Large tables, one file a month | https://data.transitlab.nyc/v1/<table>/year=YYYY/month=MM/YYYY-MM.parquet |
| Feed archives, one file a day | https://data.transitlab.nyc/v1/<table>/year=YYYY/month=MM/YYYY-MM-DD.parquet |
| The catalog, with every file’s URL | https://data.transitlab.nyc/v1/catalog.json |
| All tables as one DuckDB database | https://data.transitlab.nyc/v1/transitlab.duckdb |
Feed times are UTC. Trip times are New York local time, as the TLC publishes them. Columns say which.