Hotel Bed Places Capacity Data: Dataset Structure and Field Coverage

Hotel bed places capacity data counts sleeping capacity rather than guests. In Eurostat - Capacity of Tourist Accommodation (tour_cap_nat) the measure is one value of the accomunit dimension: BEDPL counts bed-places, sitting alongside ESTBL (establishments) and BEDRM (bedrooms), each broken out for NACE Rev. 2 groups I551 hotels, I552 holiday and other short-stay accommodation and I553 camping grounds and RV parks. A unit dimension offers NR (number) or PCH_PRE (percentage change on the previous period) for every one of those counts.

Three properties of the series shape how you use it. Coverage spans EU member states (26 reporting), EFTA members Liechtenstein, Norway and Switzerland - Iceland is not listed - plus candidate countries Montenegro, Albania and Serbia, reported at national level with EU and EFTA aggregate rows. History runs from 1990 to the latest reference year across 24,765 observations. And capacity means the maximum recorded over the reference year at establishment level, not an annual average.

National statistical institutes collect the figures through sample surveys or censuses of accommodation establishments drawn from business registers, transmit them to Eurostat via eDAMIS, and the tables are disseminated about six months after the reference year.

Of the ten primary datasets Datadory catalogs for Hotels, Resorts & Cruise Lines, only one carries the official supply census itself: tour_cap_nat, scored 9 out of 10 against a catalog-wide average of 7.81 across all 1,744 datasets. Its neighbors measure the other side of the ledger. Eurostat - Nights Spent at Tourist Accommodation (tour_occ_arn2) logs arrivals and nights spent across 286,182 observations, Hotel Booking Demand (Kaggle - 119k Portuguese Hotel Bookings) holds 119,390 individual reservations, World Bank ST.INT.ARVL covers arrivals for roughly 265 economies and the US NTTO I-94 program counts inbound visitors - none of them states how many beds exist.

Two partial substitutes cover fragments of the picture. OpenStreetMap carried 452,341 objects tagged tourism=hotel at the 21 August 2026 taginfo snapshot (255,658 nodes, 190,479 ways and 6,204 relations), and its wiki documents beds, rooms and stars attribute keys - but those tags appear on only a small minority of objects, so any country total assembled from them is a floor, not a census. CruiseMapper describes 1,531 cruise ships with per-ship specifications, yet its terms prohibit automated collection, so cruise berth capacity stays a manual-reference exercise.

The table below sorts the pool by what each source can and cannot tell you about capacity.

Verified sample pulls make the scale concrete. Greece recorded 906,866 bed-places in hotels (NACE I551) in 2024, rising to 917,212 in 2025 - a gain of 10,346 bed-places in one year, roughly 1.1%. Across all three accommodation groups (I551-I553), Greece counted 29,108 establishments in 2024, while Spain counted 58,545 in 2024 and 59,584 in 2025, adding just over 1,000 properties year on year.

One freshness boundary matters when citing: the current table already flags a 2026 reference period, but those cells registered as no-data when checked on 19 August 2026, while 2024 and 2025 were populated for most countries.

Mind the resolution mismatch in that join. Capacity stops at national level; demand descends to NUTS 2, where Veneto (ITC4) logged 7,397,598 foreign arrivals in hotels in 2024 and Stuttgart (DE11) 1,264,976 - so regional demand must be rolled up to country before dividing, or the ratio silently mixes geographies. Guest-residence segmentation only exists on the demand side too: c_resid splits every arrival figure into DOM residents, FOR non-residents and TOTAL, and there is no equivalent split of bed stock by who sleeps in it.

For ready-made intensity metrics, adjacent tables in the same tour_occ family publish annual regional occupancy rates of bed-places and bedrooms, saving you the division entirely - though those rates inherit the same national-level capacity denominator problem. The trade-offs between the two flagship tables are laid out on our capacity vs nights spent comparison page.