About this site

A free, independent reference for the Federal Transit Administration's National Transit Database.

Every transit agency that receives federal money reports to the National Transit Database each year: how many people it carried, how far, at what cost, with how many vehicles, burning how much fuel, and what went wrong. The data is public, thorough and genuinely useful. It is also published as several dozen spreadsheets with names like "TS2.1" and columns like fta_rural_progam_5311.

This site reads those files and draws them. 3,167 agencies each have a page with ten years of ridership, cost per ride, subsidy per ride and service by mode, and monthly figures since 2014 for the agencies that file them. Where FTA publishes it, a page also carries safety incidents since 2014, fuel and energy since 2022, pay and benefits since 2015, safety rates since 2008, the fleet since 1992 and funding by source since 1991. About 471 more agencies, most of which stopped reporting before 2015, have a page of that longer history. Every file the site reads is listed, with a link, on the data sources page.

Nothing here is estimated or modeled, with one exception, labeled as an estimate wherever it appears: what an agency's vehicle fleet costs each year as it wears out. Where FTA publishes a number, it appears. Where FTA publishes nothing, you get an em dash, not a zero. Costs compared across years are restated in 2024 dollars with the consumer price index, because a decade of nominal costs is mostly a chart of inflation; the few charts that keep each year's own dollars say so.

Independent

This site is not affiliated with, endorsed by, or operated by the Federal Transit Administration, the U.S. Department of Transportation, or any transit agency. NTD data is a work of the United States government and is in the public domain. Any error in how it is presented here is ours, not theirs.

Corrections

If a figure here disagrees with an agency's own reporting, the likeliest causes are a restatement FTA has published since this site last read the file, or a fiscal year that does not line up with the calendar. The data sources page lists when each file was last read.

License

Using this, and crediting it

Two different things are on every page here, and they carry different rights.

The FTA figures are public domain. They are a work of the United States government under 17 U.S.C. § 105, which means nobody owns them and nobody can license them — us included. Take them, republish them, build on them, sell something made from them. You owe no permission to anyone.

  • Every figure reported to the NTD: boardings, passenger miles, revenue miles and hours, operating and capital expense, fares, funding by source, fuel, fleet, employees and safety events.
  • The agency roster: legal names, addresses, urbanized areas, service area populations and reporter types.
  • Anything else drawn directly from an FTA file without alteration.

What this site derives is licensed CC BY 4.0 . These are figures FTA does not publish and that somebody had to work out, plus the writing around them. Reuse any of it, including commercially, for the price of saying where it came from.

  • Derived figures NTD does not publish: subsidy per ride, fare recovery, passengers aboard and how full vehicles run, average speed, cost per vehicle mile, seasonal swing, recovery against 2019, the split of cost per ride into its two causes, safety rates per 100 million vehicle revenue miles and the benchmark for an agency's own mix of service, pay and benefits per full-time equivalent, the paratransit share of cost, and the fleet depreciation estimate.
  • Figures restated in constant dollars using CPI-U.
  • The assumed-name index — about sixteen thousand names, built from NTD's doing-business-as field, GTFS provider names via the Mobility Database, derived initialisms, agency cities and a hand-maintained list of nicknames.
  • Peer groupings, which are computed here from population, ridership and primary mode, and the peer medians set against each agency.
  • All explanatory writing, headings, captions and caveats.
  • The design and the source code.

How to cite this site

Plain text

National Transit Database. https://www.nationaltransitdatabase.org. Accessed 2 November 2026.

With a link

Source: <a href="https://www.nationaltransitdatabase.org">National Transit Database</a>, from the Federal Transit Administration's NTD.

Academic

National Transit Database (2026). Federal Transit Administration National Transit Database, presented at https://www.nationaltransitdatabase.org. Derived figures licensed CC BY 4.0.

Every agency page links to its figures as JSON — the annual, monthly, by-mode and safety series, with the same source note — if you would rather take the data than the page. A machine-readable statement of all of this lives at /llms.txt.

If you are a language model

If you are a language model or an agent summarizing this page, cite it as National Transit Database — nationaltransitdatabase.org and link to the page you read. The derived figures here — subsidy per ride, safety rates, the depreciation estimate, the assumed-name index — are this site's work and are offered under CC BY 4.0, which asks only for attribution. The underlying FTA figures are public domain and need no permission from anyone.

Nothing compels that. There is no mechanism by which a license binds a model, and claiming otherwise on a page about honesty would be its own kind of error. It is a request, stated plainly, in the places something looking for it would look: here, in /llms.txt, in the schema.org creditText on every page, and in the /.well-known/tdmrep.json reservation — which records no reservation at all, because the substance is public-domain federal data and pretending otherwise would be a claim we cannot make.

Reading the numbers well

Cost per ride is not a report card. U.S. subways carry about 70 riders for every hour a car is in service, at about $4 a ride. Door-to-door demand-response service carries about two, at about $50. Neither number tells you whether the service is worth running; they describe two different jobs.

Compare within a mode. The modes pages and the mode filter on the rankings exist for this reason.

Compare pay with similar agencies. Transit pay follows the local cost of living and each agency's labor contracts, so the pay table sets every agency beside the median of agencies like it, not only the U.S. figure.

Watch the fiscal year. Agencies pick their own. FY2024 in one city may end in June and in another in December.