Back on Board
See how transit boardings changed after 2019.
Try the project ↓How to use this project
Help
How to explore
U.S. transit boardings fell sharply in 2020 and have returned unevenly. This project compares reported bus, rail, and ferry boardings with 2019; it does not count unique riders.
Start with 2025
Compare the last complete year with 2019 and see how much of the earlier boarding count returned.
Follow the monthly line
Choose all transit, bus, rail, or ferry and read each month against the same month in 2019.
Choose an agency
Select one of the 20 agencies with the most 2019 boardings and compare its annual change.
Check the season
Compare that agency month by month with 2019 so an annual total does not hide the shape of the year.
- Boarding
- One entry onto a transit vehicle; a transfer counts as another boarding, so boardings are not unique riders.
- 2019 level
- The published boarding count for the same agency, mode, or month in 2019, used here as the comparison point.
- Adjusted series
- The FTA series that includes some historical corrections and estimates for recent missing reports.
- Mode
- The broad kind of transit service; this source groups its detailed services into bus, rail, and ferry.
Back
on board
See how transit boardings changed after 2019.
A boarding is one entry onto a bus, train, or ferry. A transfer counts again, so these totals do not count unique riders.
Compare each year with 2019.
Pick a complete year. The percentage compares its boardings with 2019. Change the mode to carry the same comparison through the page.
of the 2019 boarding count
7,884,762,620 boardings in 2025, compared with 9,695,330,196 in 2019. This mode includes bus, rail, and ferry together.
Follow the monthly change.
Each point compares one month with the same month in 2019. The dashed line marks 100%, where the boarding count matches 2019.
ShowingAll transit
Latest point80.8%August 2026 versus August 2019
The same-month comparison reduces the effect of ordinary seasonal swings. It does not adjust for changed service, remote work, population, fares, or any other reason ridership changed.
Agencies changed at different rates.
These are the 20 agencies with the most boardings in 2019. Each bar shows the selected year’s count as a percentage of that agency’s own 2019 count.
Look inside the annual total.
Each pair compares one month for New York City Transit. The dark bar is 2019; the bright bar is 2025. Heights use the largest month across both years.
For New York City Transit, the monthly comparison ranges from 83.9% in Aug to 90.5% in Jan.
How to read the data and its limits
One boarding is not one rider. A person who transfers boards again and adds another trip to this measure. The totals also combine many service types inside three broad groups.
Recent values can change. FTA’s complete series adjusts some history and estimates recent missing reports. Agency service, routes, reporting, and boundaries can also change.
Most boardings returned. The pattern did not return evenly.
U.S. transit recorded 7.88 billion boardings in 2025, 81.3% of the 2019 total. Bus, rail, ferry, and the largest agencies followed different paths. This record measures use, not why people traveled or how well the system served them.
Why this exists
The question
How close were U.S. transit boardings in 2025 to their 2019 level, and how did the change differ by mode and among the 20 largest 2019 transit agencies?
What to try
Compare 2019 with 2025, change the comparison year or broad transit mode, select one of the 20 largest 2019 agencies from the agency comparison, and compare that agency's monthly pattern with 2019.
Method and limits
About the data
I used the Federal Transit Administration's Complete Monthly Ridership (with Adjustments and Estimates) dataset, discovered through the allowlisted Data.gov catalog. The National Transit Database publishes the series on data.transportation.gov and labels it U.S. public-domain material. The saved metadata identifies the Federal Transit Administration as publisher, an October 1, 2026 update, and four monthly measures. This project uses only unlinked passenger trips, the source's term for boardings. The data were retrieved October 14, 2026 and run through August 2026.
Three saved Socrata queries use the `https://data.transportation.gov/resource/8bui-9xvu.json` endpoint. The national query selects `date`, `_3_mode`, and `sum(upt)`, limits dates to January 2019 through August 2026, excludes missing trip values, and groups and orders by month and broad mode. The first agency query selects `ntd_id`, `agency`, `state`, and `sum(upt)` for calendar 2019, groups by agency ID, name, and state, orders by total trips, and keeps 20 rows. The second agency query selects those 20 IDs from January 2019 through August 2026, groups trip totals by agency, month, and broad mode, and orders by agency, month, and mode.
The four checked-in source files total 621,233 bytes. A standard-library Python reducer verifies every byte count and SHA-256, the dataset title, the FTA publisher, the public-domain label, the source update date, 92 complete national months, all three broad modes, and all 20 selected agency IDs. Repeated reductions produced the same 155,815-byte browser snapshot with SHA-256 `8d7c2bdc6e16c79ee725a57d049d8562b50c821c9522f7da3bac5c2f604a18db`. The page imports that snapshot and makes no runtime request.
Annual values sum the 12 published monthly boarding counts for each calendar year. A percentage of the 2019 level divides a later count by the matching 2019 count. The monthly line compares each later month with the same calendar month in 2019, which controls for the ordinary month-of-year pattern but does not adjust for service changes or other conditions. The agency view is intentionally limited to the 20 largest agencies by 2019 boardings; it does not rank every U.S. transit provider.
A boarding is not a person or a complete journey: one rider can board several vehicles and be counted several times. The complete series includes FTA adjustments and estimates for recent missing reports, and later releases can revise it. Agencies can change service, reporting, routes, or organizational boundaries. The data do not explain why use changed and cannot establish service quality, access, fare revenue, commuting behavior, or cause.
Skills used
- Data analysis
- Data engineering
- Statistical reasoning
- Visualization
- Information design
- Interaction design
- Visual design
- Creative coding
- Product thinking
- Editorial storytelling
- Accessibility
- Performance
- Systems design
What the data shows
- The series records 9.70 billion U.S. transit boardings in 2019, 4.51 billion in 2020, and 7.88 billion in 2025. The 2025 total is 81.3% of the 2019 total.
- In 2025, bus boardings are 83.6% of their 2019 level, rail is 79.1%, and ferry is 85.2%. These are broad national groups that combine many different services and places.
- Among the 20 agencies with the most 2019 boardings, 2025 ranges from 99.4% of the 2019 level for MTA Bus Company to 48.5% for Bay Area Rapid Transit. That comparison reflects agency boundaries and reported boardings, not service quality or the share of residents using transit.
- August 2026 boardings equal 80.8% of August 2019 in this release. The 2026 year is incomplete, includes possible estimates, and can be revised, so the project keeps it in the monthly view rather than treating it as a finished annual result.