Datadory notebook
Hotels, Resorts & Cruise Lines Data for Market Researchers
Market researchers use hotels, resorts and cruise lines data to size accommodation demand, benchmark room supply and track cruise deployment from official statistics: Eurostat's regional nights and capacity tables, World Bank arrivals and receipts for roughly 265 economies, US NTTO I-94 inbound files and CruiseMapper's 1,531-ship fleet database. Nine of the ten primary datasets are free.
1,744 datasets. Pick your catch.
What does a hotels, resorts and cruise lines brief actually ask you to deliver?
The sources fall into three families. Official tourism statistics - Eurostat's capacity and nights-spent tables, the World Bank arrivals and receipts indicators, the US NTTO I-94 program and UK ONS travel and tourism - set the market-sizing frame. Row-level operational bookings (the Kaggle Hotel Booking Demand file) supply guest behavior. Live operational trackers - OpenStreetMap points of interest and CruiseMapper's fleet database - describe physical supply and deployment. In practice most studies pull one source from each family, document all three in the methodology appendix, and refresh annually. Start from the hotels-resorts-cruise-lines data hub and the hotels, resorts and cruise lines data for market researchers page for the stacked view.
Which datasets anchor the market-sizing layer?
Two Eurostat tables carry most European sizing work. Eurostat – Capacity of Tourist Accommodation (tour_cap_nat) is the only public census of accommodation supply: annual counts of establishments, bedrooms and bed-places for hotels (NACE I551), holiday and other short-stay accommodation (I552) and camping grounds and RV parks (I553). It covers 26 reporting EU member states plus EFTA countries (Liechtenstein, Norway, Switzerland - Iceland is not listed) and candidate countries Montenegro, Albania and Serbia, runs from 1990 onward with a 2026 reference flag already present, and holds 24,765 observations under DOI 10.2908/TOUR_CAP_NAT. That bed-place count is the denominator researchers quote when a client asks whether a market is under-roomed.
How do you build a destination demand indicator step by step?
The fastest route from raw tables to a deck-ready occupancy leading indicator is a five-step join:
- Pull nights spent and arrivals from tour_occ_arn2 for your target NUTS 2 regions via the JSON-stat API - no authentication, and the flat TSV export runs to roughly 16,400 rows.
- Split the resident versus non-resident series so international demand is isolated from domestic travel.
- Overlay the US inbound mix from the US NTTO I-94 International Visitor Arrivals Program - monthly country-of-residence Excel workbooks from 2000 to present, built from DHS/CBP ADIS records covering over 355 million travelers, broken out by visa type, port of entry and first intended address.
- Convert volume into money by dividing World Bank receipts by arrivals: ST.INT.RCPT.CD current-US$ receipts over ST.INT.ARVL arrivals yields revenue per visitor across roughly 265 economies. Mind the gaps - populated observations run effectively 1995 to 2020, with 206 economies reporting arrivals in 1995, 223 in 2019 and 132 in 2020, while receipts come from 130 to 175 economies a year.
- Sanity-check totals against the Statistics Explained summary workbooks before the chart ships.
The result is a defensible demand-and-yield table per market that cites national statistical offices rather than a paywalled estimate - the difference between a methodology appendix that survives client scrutiny and one that does not.
Where does booking-level behavioral evidence come from?
Scope it honestly in the sampling frame: two properties in one country is a methods testbed, not a national read. It is the standard choice for prototyping cancellation and no-show models, demonstrating ADR seasonality or stress-testing a revenue-management storyline before commissioning proprietary data - and because it is static, note the fixed 2015-2017 window in any forecast that leans on it.
Can you map competitive supply without buying POI data?
For site-selection and density questions, OpenStreetMap substitutes for commercial point-of-interest vendors. OpenStreetMap Hotels via Overpass Turbo queries the live global database for every object tagged tourism=hotel - 452,341 objects in the taginfo snapshot dated 2026-08-21 (255,658 nodes, 190,479 ways, 6,204 relations) - alongside 329 tourism=resort entries and hundreds of thousands of hostel and guest-house objects. Each hit returns geocoded coordinates with name, stars, rooms, beds, address, phone and website tags, exportable as JSON, GeoJSON or CSV through the Overpass API in real time, with historical states available via date filters since Overpass v0.7.55.
Can you cite these sources in published client deliverables?
The two edges of the envelope: OpenStreetMap's commercial delivery terms 1.0 is share-alike, so derived databases inherit the obligation, and CruiseMapper's terms prohibit automated collection and commercial exploitation without written consent while permitting republishing only with accreditation to CruiseMapper.com - governed by Bulgarian law, with all accuracy warranties disclaimed. Cite the nine free sources freely; handle the tenth manually or by negotiation. The curated open tier is browsable on the free hotels, resorts and cruise lines datasets list.
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
Eurostat - Capacity of Tourist Accommodation (tour_cap_nat)
Eurostat – Nights Spent at Tourist Accommodation (tour_occ_arn2)
Eurostat Tourism Statistics Explained Hub
Indicator · geo · time …+1 more
US NTTO I-94 International Visitor Arrivals Program
World Bank - International Tourism Arrivals (ST.INT.ARVL)
Hotel Booking Demand (Kaggle – 119k Portuguese Hotel Bookings)
hotel · lead_time · adults …+14 more
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sample