Yellow taxi trips

Every yellow cab trip in New York City since January 2023, from the Taxi and Limousine Commission, with readable names and zones.

Rows: 71,370,274
Size: 1.3 GB
Covers: January 2023 onward; the TLC publishes about two months late
Updated: Monthly
Source: NYC TLC trip record data
License: NYC open data terms

Fileshttps://data.transitlab.nyc/v1/taxi_yellow_trips/year=YYYY/month=MM/YYYY-MM.parquetOne file a month, 18 so far. In the database as transitlab.taxi_yellow_trips.

Rows and size from catalog.json, 2026-10-06.

What a row is

One taxi trip, from meter on to meter off. Pickup and drop-off are taxi zones, not points: 263 areas across the city.

Each file is the TLC’s file for that month with a few changes, so every month has the same columns:

Readable names. Columns are renamed to one style across yellow, green and app trips (pickup_datetime, trip_miles, tip_amount), and codes are spelled out (payment_type is cash, not 2).
Zones added. Each trip has the name and borough of its pickup and drop-off taxi zone, so you rarely need the lookup table.
Light cleaning. We keep trips that start in the file’s month, end after they start, and last under a day. Nothing else is dropped, so outliers remain.

Query it

-- Trips and average speed by hour, August 2026, trips within Manhattan
SELECT hour(pickup_datetime) AS hour,
       count(*) AS trips,
       round(sum(trip_miles) / sum(trip_minutes / 60), 1) AS mph
FROM transitlab.taxi_yellow_trips
WHERE pickup_datetime >= '2026-08-01' AND pickup_datetime < '2026-09-01'
  AND pickup_borough = 'Manhattan' AND dropoff_borough = 'Manhattan'
  AND trip_miles BETWEEN 0.3 AND 30 AND trip_minutes BETWEEN 2 AND 90
GROUP BY hour
ORDER BY hour;

Or read one month straight from its file:

SELECT payment_type, count(*) AS trips, round(avg(tip_amount), 2) AS avg_tip
FROM 'https://data.transitlab.nyc/v1/taxi_yellow_trips/year=2026/month=08/2026-08.parquet'
GROUP BY payment_type
ORDER BY trips DESC;

Columns

vendor_idsmallint

Meter vendor: 1 Creative Mobile, 2 VeriFone, 6 Myle, 7 Helix.

pickup_datetimetimestamp

Meter on, New York local time.

dropoff_datetimetimestamp

Meter off, New York local time.

passenger_countsmallint

Passengers, as the driver entered it.

pickup_zone_idsmallint

TLC taxi zone of the pickup (1 to 263).

pickup_zonevarchar

Name of the pickup zone, e.g. Midtown Center.

pickup_boroughvarchar

Borough of the pickup zone, or EWR, Unknown, N/A.

dropoff_zone_idsmallint

TLC taxi zone of the drop-off.

dropoff_zonevarchar

Name of the drop-off zone.

dropoff_boroughvarchar

Borough of the drop-off zone.

trip_milesdouble

Distance in miles.

trip_minutesdouble

Minutes from pickup to drop-off.

rate_codevarchar

standard, jfk, newark, nassau_westchester, negotiated, group or unknown.

store_and_fwdboolean

True if the meter held the record before sending it.

payment_typevarchar

credit_card, cash, no_charge, dispute, flex_fare, voided or unknown.

fare_amountdouble

Time and distance fare, in dollars.

extradouble

Rush hour and overnight extras.

mta_taxdouble

MTA tax.

tip_amountdouble

Card tips only; cash tips aren't recorded.

tolls_amountdouble

Tolls paid.

improvement_surchargedouble

Improvement surcharge.

congestion_surchargedouble

New York State congestion surcharge.

airport_feedouble

Fee for pickups at LaGuardia and JFK.

cbd_congestion_feedouble

MTA congestion pricing toll, from 2025-01-05. Null before.

total_amountdouble

Total charged, cash tips excluded.

request_sourcevarchar

How the trip was requested. New in 2026 files; null before.

Good to know

The TLC publishes about two months late. August appears in October.
Trim before you average. Expect trips of zero miles and odd fares. Our taxi study keeps trips of 2 to 90 minutes, 0.3 to 30 miles and 1 to 60 mph.
Times are New York local time, with no zone attached.
Cash tips aren’t recorded, so tip_amount means something only for card trips.
Terms: NYC public data. Credit the NYC Taxi and Limousine Commission.