US NTTO I-94 International Visitor Arrivals Program

One market, counted properly. The International Visitor Arrivals Program is the National Travel and Tourism Office's (NTTO) official count of visitors arriving in the United States, produced with the Department of Homeland Security and US Customs and Border Protection from ADIS I-94 travel histories spanning over 355 million unique travelers. An arrival means a visitor staying one night or longer under covered visa types: overseas travelers of every mode of transport, plus Mexico air and sea travelers. Country of residence is derived overwhelmingly from airline APIS manifests - about 99.89 percent of the country-of-residence data - so the origin attribution rests on flight manifests rather than on anyone's guess at a border booth.

The payoff is dimension. Where the widest international series carry a single annual national total per economy, this program answers for one market with cuts that analysts actually ask for: monthly and annual country-of-residence series running 2000 to present, a ranked top twenty of overseas tourist-generating countries each month (the United Kingdom led July 2026's preliminary table, ahead of India, Japan, Brazil and China), visa type across nineteen codes, air/sea/land mode of transport, seven age bands, first intended US state, and roughly forty named ports of entry. If your question names America - how large is the UK inbound market this July versus last July, which visa classes recovered fastest, what share of arrivals clears Miami - this is the count the US travel industry quotes.

The tables arrive pre-formatted: twenty-plus lettered tables (A through Q+) per reporting period, each holding one cut of the same underlying count. Verified rows read like this:

# US NTTO I-94 -- arrivals tables, July 2026 (preliminary)
world_region_or_cor   rank   arrivals   ytd_share   yoy_vs_2025
United Kingdom        1      1264976    0.09        4.2

# dimension cuts carried alongside the country tables
visa_code   mode       age_group   mean_age   first_intended_state   port_of_entry
B-2         Air Mode   25 - 34     44.6       Florida                Miami, FL

The headline table ranks the top twenty overseas tourist-generating markets; July 2026's preliminary ordering runs United Kingdom, India, Japan, Brazil, China (PRC), Germany, France, South Korea, Colombia, Italy. Below it sit the splits: visa families (business B-1/WB/E-1/E-2, pleasure B-2/WT, student F-1/M-1 among the nineteen codes), mode of transportation with its own 'Not Reported' bucket rather than silent drops, age bands running 0-17 through 65-or-over with mean and median age per world region, first intended address by US state and territory, and port-of-entry tables kept in two versions - all modes and air-only - because a land crossing into Blaine, WA and a transatlantic arrival into Miami are different stories about the same country.

Twelve fields carry most analyses, and each definition below describes exactly what the delivered column holds, worked against a real value. Two boundaries deserve a read before you model anything: Canadian traveler figures come from Statistics Canada's International Travel Survey and Mexican land-border arrivals from INEGI/Banco de Mexico partner series, so the program proper counts overseas travelers of all modes plus Mexico air and sea. And 'first intended address' is stated intent at admission - a Florida entry does not guarantee a Florida bed, so treat state as directional demand rather than room-night truth.

Three chips summarize the footprint:

• Geography: the United States as destination, with origin markets worldwide - individual country-of-residence rows plus world-region groupings and a top-twenty overseas ranking, alongside roughly forty named US ports of entry (all-modes and air-only lists) and all US states and territories as first intended address.
• Time frame: country-of-residence series run 2000 to present in both monthly and annual form - a quarter-century of continuous history that carries pre-2020 peaks, the collapse, and the recovery in one comparable frame. Figures arrive in preliminary and final vintages, and preliminary values stay open to revision for up to 36 months, so pin the vintage you build on.
• Granularity: monthly and annual report-table grain, keyed by country or region crossed with visa type, mode of transportation, age band, US state and port of entry. This is structured tabular detail, not microdata: no individual traveler records exist here, and none are needed to size a market.

Against the wider Datadory catalog - average quality score 7.81 across all 1,744 datasets - this slice scores 9/10, carried by verified field documentation and a twenty-six-year continuous series under one unchanged counting concept.

API, files, or your warehouse. Daily, weekly, or hourly.

You choose the channel and the cadence; the field dictionary above travels unchanged across all three. Bulk files suit overnight warehouse loads, structured payloads suit event-driven pipelines, and the period-by-market tables land in CSV-, spreadsheet- and SQL-friendly shapes for analysts who live in a BI tool. Because the table structure never changes between periods, switching cadence later is a settings conversation, not a re-integration project - and the sample you request lands pre-cut to the markets, months and dimensions you name.

Ranked by how directly the program answers their day job:

1. Market researchers & consultants. Official, consistently-defined counts of who arrives in the world's largest inbound travel market, by month and source country, are the baseline any US market-sizing deck gets audited against - and the top-twenty ranking hands you the league table pre-built.
2. Data scientists & ML engineers. Twenty-six years of monthly history with visa, mode, age and geographic dimensions is training material for demand forecasting with actual regime variety - growth, shock, recovery - and the report-table grain drops into a dataframe without scraping work.
3. Investors & quants. Monthly inbound flows by source market are a high-frequency demand signal for lodging and cruise exposure; join country-of-residence trends to revenue lines and 'travel recovery' becomes a number with a date on it.
4. Route development & aviation analysts. Port-of-entry tables in all-modes and air-only versions, crossed with origin country and visa class, are the raw material for gateway performance, new-route cases and connectivity arguments.
5. Hotel, DMO & cruise commercial teams. Mode-of-travel splits isolate sea arrivals for itinerary planning, state-of-intended-address shows which gateways feed which regions, and source-market rankings set where sales offices and campaigns go next.

One honest limit: this counts arrivals, not behavior. What a visitor paid, where they actually slept, or whether they canceled belongs to property-level datasets - pair them when both halves of the journey matter.

Notes that pair well with this page:

• Hotels, Resorts & Cruise Lines data hub - the pooled view of the industry, from booking records to economic indicators (hub).
• World Bank - International Tourism Receipts (ST.INT.RCPT.CD) - the money behind the headcounts; US arrivals here joined to receipts there turns traffic into yield (receipts).
• World Bank - International Tourism Arrivals (ST.INT.ARVL) - the global breadth play: one annual total for ~265 economies where this program holds deep monthly detail for one market (arrivals).
• Hotel Booking Demand (Kaggle - 119k Portuguese Hotel Bookings) - the other half of the telescope: 32 behavioral fields per reservation inside the property, versus census-style counts at the frontier (bookings).
• Head-to-head: Hotel Booking Demand vs US NTTO I-94 - a dedicated page weighing per-booking behavioral depth against official national scale (comparison).
• Industry guide - how hospitality and travel datasets fit together across the catalog (hotels-resorts-cruise-lines data guide).