How we measure TSA wait times: data & methodology

Two tiers of data, one honest label each

We do not average, blend, or launder the two sources into one number. Each airport page shows exactly one kind of wait, labeled for what it is.

Tier A — live official checkpoint feeds

Where an airport publishes its own checkpoint wait feed, we read that feed directly and show its number with a Live feed pill. 11 of our core 32 airports currently qualify, including Atlanta (ATL), Los Angeles (LAX), San Francisco (SFO), and Miami (MIA). The label is conditional, not decorative: a number only reads "Live" when the real feed actually answered on the last poll. If a feed goes quiet, the page degrades to a clearly labeled prediction instead of freezing a stale number under a live badge.

Tier B — the flight-schedule demand model

Every other airport gets a Predicted wait from our own model. Its inputs, in order of influence:

• Scheduled flight volume (FIDS data). The day's departure schedule is the primary demand signal — more seats leaving in the next few hours means more people in line now.
• Hour-of-day curves. Checkpoint demand follows a reliable double hump — an overnight lull, a morning peak around 8 AM, and an evening peak around 6 PM — which shapes the schedule signal into a wait estimate.
• Day-of-week patterns. Typical weekday-vs-weekend checkpoint behavior adjusts the curve; Sunday and Friday run hotter than Tuesday.
• Holiday demand. Peak travel dates enter mostly through the flight schedule itself — airlines add capacity around holidays, and the model sees that added volume directly rather than relying on a hard-coded holiday calendar.

If schedule data is briefly unavailable, the model falls back to the typical time-of-day curve alone — still labeled as a prediction. We never dress a prediction up as a live reading.

Coverage: which airports are live vs modeled

The table below is the coverage snapshot for our core 32 airports — the large U.S. hubs where we maintain the deepest data. Beyond this core set, the site tracks security waits at 150+ U.S. airports via the Tier-B model. Airport codes link to each live wait page.

Snapshot as of July 2026. Feeds get added as airports publish them; the on-page label at any airport is always the current truth.

Refresh cadence

• Live pages: the feed is re-polled about every 120 seconds while the page is open, and a timestamp under the departures board shows when the number last changed — no guessing about staleness.
• Predicted pages: estimates recompute through the day as the flight schedule progresses, so the answer at 6 AM differs from the answer at 2 PM.
• Published datasets (/data pages): dated snapshots — each states its compute or access date and is revised on the page, with the updated date refreshed, when we confirm a correction.

Known limitations

Honest numbers require honest caveats. The ones that matter most:

• Modeled waits are estimates, not measurements. Tier-B figures — and any published table built from them — are first-party modeled estimates from our forecasting model, never TSA or government data. Treat per-airport model figures as "our model's typical-day estimate," not a measured ranking of airports.
• Standard lanes only. Our waits describe the general screening line. PreCheck lanes run much shorter — TSA's own service standard is under 30 minutes in standard lanes and under 10 minutes in PreCheck lanes.
• Live feeds are only as good as their source. We relay official airport numbers; if an airport's feed lags reality, so do we. The two-minute poll bounds staleness on our side but cannot correct the source.
• Disruptions are out of scope. The model captures typical demand, not one-off shocks — a staffing incident, weather closure, or equipment outage can push a real line past any schedule-driven forecast.
• Snapshot datasets can lag the live model. A published /data snapshot reflects its stated compute date; the live site keeps moving after the snapshot is taken.

Full method note for our published model snapshots: All wait figures are FIRST-PARTY MODELED ESTIMATES produced by tsawaittimes.app’s own forecasting model (the same Tier-B model that powers the live site), computed 2026-07-03 for the launch set of 32 large U.S. airports. They are not TSA or government measurements. For this snapshot every airport was computed via the model’s deterministic typical-day curve (an overnight lull with ~8 AM and ~6 PM peaks) scaled per airport, clamped to 4–75 minutes; the flight-schedule-density input was unavailable at run time. The national hour-by-hour shape is the model’s real typical-day curve, but per-airport differences come from the model’s seeded scaling — treat per-airport rows as “our model’s typical-day estimate,” never as a measured ranking of airports. Averages weight operating hours (4 AM–10 PM) at 1.0 and overnight hours at 0.25. Standard lanes only; 11 of the 32 airports also have live Tier-A checkpoint feeds (marked “live”). No day-of-week or seasonal term is included. Values rounded to whole minutes.

Independence

tsawaittimes.app is an independent travel utility. We are not affiliated with the TSA, any airport authority, or any airline — and no airport or agency pays for placement or influences a number. Who we are and how the written guides are checked lives on the About page and in How it works. For what the waits mean for your own departure timing, start with how early to arrive at the airport.

For journalists & researchers

We built the /data section to be quoted. Three ground rules make that easy for everyone:

• Attribution.Credit "tsawaittimes.app" with a link to the page you used. The standing description is: "Data: tsawaittimes.app, which tracks security waits at 150+ U.S. airports."
• Labeling.Figures from our model must be described as modeled estimates from tsawaittimes.app's forecasting model — not as TSA, airport, or government measurements. Live-feed readings may be described as official airport feed data relayed by tsawaittimes.app.
• Data on request. Underlying tables (CSV/JSON) behind any /data page are available to journalists and researchers on request, including method notes and source lists.

Cite or share this data

Plain-text citation for this page:

Source: tsawaittimes.app — How we measure TSA wait times: data & methodology, 2026

Charts and tables from our /data pages are free to republish with a link to the original page, under CC BY 4.0. No permission needed — attribution is the whole license.

What does the Live label mean on tsawaittimes.app?

Live means the number was read directly from the airport's own official checkpoint feed within roughly the last two minutes — currently possible at 17 major U.S. airports, including Atlanta (ATL), Los Angeles (LAX), Charlotte (CLT), Houston (IAH), and Washington National (DCA). A wait only wears the Live feed pill when the real feed actually answered; if the feed goes quiet, the page falls back to a clearly labeled prediction rather than showing a stale number as live.

How are predicted TSA wait times calculated?

Predicted waits come from our own demand model. Its primary input is that day's scheduled departure volume (FIDS flight-schedule data), shaped by typical checkpoint curves for the hour of day and day of week — an overnight lull with morning (~8 AM) and evening (~6 PM) peaks. Holiday surges enter mainly through the flight schedule itself, since airlines add capacity on peak dates. If schedule data is briefly unavailable, the model falls back to the typical time-of-day curve alone. Every predicted number is labeled as a prediction, never as a live reading.

How often are the wait times updated?

Live airport pages re-poll the official feed about every 120 seconds while you have the page open, and a timestamp under the board shows when the number last changed. Predicted waits recompute continuously through the day as the flight schedule progresses, so a quiet 2 PM and a slammed 6 AM get different answers. Published datasets on our /data pages are dated snapshots and state their compute date.

How should I cite tsawaittimes.app data?

Use the line: "Data: tsawaittimes.app, which tracks security waits at 150+ U.S. airports." For a specific page, cite it as "Source: tsawaittimes.app — <page title>, 2026" with a link to the page. Our charts and tables are free to republish with attribution under CC BY 4.0. Modeled figures should be described as estimates from tsawaittimes.app's forecasting model, not as TSA or government measurements.