How to Use NYC TLC Trip Record Data (Driver's Guide)

In short: TLC publishes every completed taxi and app trip as monthly Parquet files on its Trip Record Data page: one file each for yellow taxis, green cabs, FHVs and high-volume FHVs (Uber and Lyft). On October 12, 2026 the newest month was August 2026, because TLC publishes "typically with a two-month delay." If you only need monthly totals by zone or industry, skip the big files and use two NYC Open Data tables: c5iv-bn4s and v6kb-cqej.

This guide is for drivers and anyone curious: what each file holds, the fields that matter for money, and the mistakes that make people publish wrong numbers. All the trip statistics on TLC Help come from these same files. For a one-line definition, see trip record data.

The four record types

TLC warns that yellow and green data "was not created by the TLC, and TLC makes no representations as to the accuracy." FHV records "are generated from the FHV Trip Record submissions made by bases, so we cannot guarantee or confirm their accuracy or completeness." Since 2025 the yellow, green and HVFHV files also carry a cbd_congestion_fee column for congestion pricing (TLC).

Field guide: the columns drivers care about

High-volume FHV (Uber/Lyft) — from TLC's HVFHV dictionary, dated March 18, 2025:

• hvfhs_license_num — the company. HV0003 = Uber, HV0005 = Lyft (HV0002 Juno and HV0004 Via appear only in old files).
• request_datetime, pickup_datetime, dropoff_datetime — request, pickup and drop-off times. on_scene_datetime is filled for accessible vehicles only.
• PULocationID, DOLocationID — pickup and drop-off taxi zone numbers.
• trip_miles, trip_time (in seconds) — the passenger part of the trip only.
• base_passenger_fare — the fare "before tolls, tips, taxes, and fees."
• driver_pay — "Total driver pay (not including tolls or tips and net of commission, surcharges, or taxes)."
• tips, tolls, bcf (Black Car Fund), sales_tax, congestion_surcharge, airport_fee, cbd_congestion_fee.
• Flags: shared_request_flag, shared_match_flag, wav_request_flag, wav_match_flag, access_a_ride_flag.

Yellow taxi — from the yellow dictionary:

• RatecodeID — 1 standard, 2 JFK, 3 Newark, 4 Nassau or Westchester, 5 negotiated, 6 group ride, 99 unknown.
• payment_type — 0 Flex Fare trip, 1 credit card, 2 cash, 3 no charge, 4 dispute, 5 unknown, 6 voided.
• fare_amount — "the time-and-distance fare calculated by the meter"; extra, mta_tax, improvement_surcharge, congestion_surcharge, airport_fee, cbd_congestion_fee are separate.
• tip_amount — "automatically populated for credit card tips. Cash tips are not included."
• trip_distance — miles "reported by the taximeter."

Reading one app trip, column by column

Imagine one row from an HVFHV file. hvfhs_license_num says HV0003, so it is an Uber trip. PULocationID 168 and DOLocationID 138 mean it started in Mott Haven/Port Morris (Bronx) and ended at LaGuardia (look both up in the zone table). trip_miles 7.9 and trip_time 1,260 mean 7.9 miles and 21 minutes with the passenger. base_passenger_fare is the price of the ride itself; airport_fee, sales_tax, bcf and congestion_surcharge are separate charges collected on the trip. driver_pay is the driver's pay for the trip, and tips and tolls are recorded in their own columns because driver_pay leaves them out. So to see everything tied to the driver on that trip you add driver_pay and tips, and look at tolls separately.

For a yellow taxi row, the driver's side is less direct. The meter fare, the extras and card tips are there, but what the driver keeps depends on the lease, and cash tips are invisible.

How to answer a question with the raw files

A minimal DuckDB query for the example above looks like this (file names as TLC publishes them):

SELECT median(driver_pay) AS median_pay, count(*) AS trips
FROM 'fhvhv_tripdata_2026-08.parquet' t
JOIN 'taxi_zone_lookup.csv' z ON z.LocationID = t.PULocationID
WHERE z.Borough = 'Bronx' AND trip_miles > 0;

Filtering by borough this way is how a page like our Bronx page can describe local trips. Note what the files cannot do: the tolls column gives a dollar amount, not the crossing, so tying a trip to a bridge such as the Bronx–Whitestone means guessing from the pickup and drop-off zones.

Open Data shortcuts (no big files)

If you only need totals, two NYC Open Data tables already did the counting:

• Pickups and Drop-offs by Taxi Zone and Industry (c5iv-bn4s) — monthly pickup and drop-off counts per taxi zone for yellow, green and high-volume FHV. The city describes it as made "for a wider audience, including those without the technical expertise to work with raw trip record data." You can filter and export it in the browser.
• TLC Industry Indicators (v6kb-cqej, also a CSV) — every industry, every month since 2010: trips per day, farebox per day (taxis only), unique drivers and vehicles, days and hours on the road, minutes per trip and percent paid by card. It is the best long-run series. Its "hours per day" counts hours in which a trip was recorded, not logged-in time; see what TLC data says about driver hours.

Common mistakes

Two more: drivers and vehicles overlap across industries in the indicators (one TLC license covers taxi and FHV work), so never add driver counts across lines. And yellow files gained a request_source column in 2026 that the March 2025 dictionary does not define; count it, but don't guess what its codes mean.

Example questions the data answers

Using September 2025 – August 2026 files (our aggregation; methods on each page):

• What does a typical app trip pay? → Uber/Lyft pay per trip, mile and minute.
• How much do riders tip? → yellow taxi card tips and app tipping.
• How long are trips? → trip length distribution.
• How do fares compare? → taxi vs Uber fares, yellow fare per trip and what's inside a taxi fare.
• The big picture → NYC taxi and Uber trip statistics.

What the data does and doesn't mean

Trip records describe completed trips. They do not show time spent waiting for a dispatch, miles driven empty, cancelled requests, bonuses paid outside a trip, or your own costs. They describe the market, not one driver. Your own pay history is on your app statements, and your TLC license records are in TLC UP. This guide explains public data; it is not legal or tax advice. More money topics: money hub.