Methodology: how Passenger Index measures US airport traffic

Sources, definitions and caveats · public data through September 30, 2026

Passenger Index tracks US air travel demand airport by airport by combining two federal datasets: departing passengers screened at TSA security checkpoints and international passengers processed by US Customs and Border Protection on arrival. It covers 458 airports, 45 of them with international arrivals, with daily data from 2015.

Sources

TSA checkpoint throughput

TSA's FOIA reading room publishes throughput by airport, checkpoint and hour (tsa.gov/foia/readingroom). Passenger Index extracts every release, sums checkpoints to airport-day and keeps one value per airport-day where releases overlap. History from January 4, 2015.

CBP international arrivals

CBP Airport Wait Times (awt.cbp.gov): hourly passengers processed at primary inspection, US citizens and non-US citizens, for 45 airports. History from May 3, 2023; CBP keeps a rolling ~3-year window.

National TSA volumes

The free dashboard also shows TSA's national daily checkpoint count (tsa.gov/travel/passenger-volumes), which has no airport split.

License

Source data is US federal public domain (17 U.S.C. §105). Passenger Index's public summaries are licensed CC BY 4.0.

Definitions

Departing passengers screened (TSA)

Passengers screened at the airport's TSA checkpoints. It counts every screened departure, domestic and international flights alike: TSA does not record destination, so it cannot be split into domestic and international.

International arrivals (CBP)

Passengers processed by CBP at primary inspection on arrival (US citizens plus non-US citizens), at the 45 CBP-monitored airports only.

Why they are not added

One series counts departures and the other arrivals, so their sum is not a meaningful passenger count. They are shown side by side.

Trailing 12 months

The 365 days ending on the as-of date. If some days are missing, the sum is scaled to 365 days (shown on the page); under half the days, no figure is published.

Year-over-year (YoY)

The same window 364 days earlier, so weekdays line up, summed over days present in both years only. A missing data week never reads as a change in demand.

Hub class

FAA-style classes by share of national TSA checkpoint throughput over the trailing 12 months: large hub ≥1%, medium 0.25-1%, small 0.05-0.25%, otherwise non-hub. The FAA's official classes use calendar-year enplanements.

Rank

Position among all 458 airports by trailing-12-month departing passengers screened.

Caveats

• Not airport-reported totals. Airports report enplaned plus deplaned passengers. TSA screens departing passengers once; domestic arrivals and connecting passengers who stay airside are not screened again. Connecting hubs such as ATL, CLT, DFW and DEN show far fewer passengers here than in airport statistics.
• Preclearance. Travelers who clear US immigration abroad (CBP preclearance in Canada, Ireland, the Bahamas, Bermuda, Aruba, the UAE and elsewhere) land as domestic passengers and are not in the CBP count at the US airport.
• CBP covers 45 airports. CBP Airport Wait Times lists 45 airports; other international gateways have no international-arrivals figure.
• Gaps are not filled. TSA FOIA releases have missing weeks and CBP has outages. Missing days are never interpolated; windows are scaled from the days present and YoY uses matched days only. Days with a checkpoint missing are handled the same way (incomplete days).
• Checkpoint mapping. TSA reports by checkpoint; Passenger Index sums checkpoints to the airport TSA assigns them to.
• Revisions. CBP revises recent days; TSA occasionally reissues releases. Figures are rebuilt from source on every run.

Incomplete days in TSA's data

TSA reports passengers separately for each security checkpoint. On some days a checkpoint that normally reports is missing from TSA's files. Adding up only the checkpoints that are there would understate that airport's day, sometimes by most of its traffic, and distort every total and year-over-year change built on it.

How a day is identified as incomplete. An airport's day is incomplete when a checkpoint that is expected that day has no data. A checkpoint is expected on a day when all three hold:

1. it reported on at least one of the 7 days before,
2. it reported on at least one of the 7 days after, so a checkpoint that opens or closes for good is never treated as missing (for the most recent 7 days this test waits for the next week's data, so those days are provisional), and
3. it normally handles at least 3% of the airport's passengers, measured over the 29 days around that day, so a small checkpoint that closes on quiet days does not set the whole day aside.

What happens to an incomplete day. It is treated like a day TSA never reported. Nothing is estimated or filled in.

• 7-day, 30-day and 12-month totals, monthly, weekly and annual figures and daily averages use the complete days, scaled to the full period, as for any missing day; a period needs at least half its days complete.
• Year-over-year changes compare only days that are complete in both years; with fewer than half the days matched, the change is shown as n/a.
• Daily charts show a gap on an incomplete day.
• Where a period's change is affected, the figure is marked; hover or tap it to see how many days were left out and the change as reported, including them. Downloads carry the same information in their own columns.

Which data and periods. Every airport-day of TSA checkpoint data, from Passenger Index's first day in 2015 to the latest, is checked on every rebuild. It does not apply to TSA's national daily total, which TSA publishes as one number, or to CBP international arrivals, which are reported per airport.

Across all Passenger Index data, 4,156 airport-days at 95 airports are set aside this way, 1.9% of days at airports with more than one checkpoint:

Airports with the most days set aside: ACY (583), PIE (455), LIH (273), LGA (266), GRK (206), VPS (164), BWI (148), HNL (147), OGG (146), IAH (139).

Public pages vs Premium

Passenger Index Premium

How to cite

Passenger Index, "US airport passenger traffic", data through September 30, 2026. https://passengerindex.com/web/methodology/. Based on TSA FOIA checkpoint throughput and CBP Airport Wait Times.

Frequently asked questions

What does Passenger Index measure?

Passenger Index measures US air travel demand at the airport level from two federal sources: departing passengers screened at TSA checkpoints (from TSA's FOIA throughput releases) and arriving international passengers processed by US Customs and Border Protection (from CBP Airport Wait Times).

Is Passenger Index's figure the same as an airport's reported passenger count?

No. Airports report enplaned plus deplaned passengers, including connections. Passenger Index ranks airports by departing passengers screened at TSA checkpoints, which leaves out arrivals and airside connections, so it is lower - by much more at connecting hubs.

Can Passenger Index split passengers into domestic and international?

No. TSA screening does not record a passenger's destination, and CBP counts international arrivals only, so there is no international-departures figure to subtract. Departing passengers screened and international arrivals are shown side by side; they count opposite directions and are never added.

Which airports have international arrival data?

The 45 airports in CBP's Airport Wait Times data: ATL, AUS, BNA, BOS, BWI, CLT, CVG, DEN, DFW, DTW, EWR, FAT, FLL, GUM, HNL, IAD, IAH, JFK, LAS, LAX, MCO, MDW, MIA, MSP, OAK, ONT, ORD, PBI, PDX, PHL, PHX, PVD, RDU, SAN, SAT, SEA, SFB, SFO, SJC, SJU, SLC, SMF, SNA, STL, TPA. Other airports have no international-arrivals figure.

How current is the data?

Public pages show the 12 months to September 30, 2026. Premium data runs to October 3, 2026, the latest day in TSA's FOIA releases, which TSA posts weekly, about a week after the fact.

Can I use these figures?

Yes. Public summaries are licensed CC BY 4.0: cite Passenger Index and link to the page. The underlying TSA and CBP data is US federal public domain.

Airports

All 458 airports ranked by 12-month passengers →