Somebody Photographed Prague's Hotels Address by Address

Czechia is the best-photographed country in our entire catalogue: 889 pictures on 897 rows, 99.1%, a higher rate than Spain, Austria, Britain, Japan or the United States. It is also one of the emptiest — no telephone, no description, no star grade anywhere in it. But the pictures are not anonymous. Their filenames read “Praha Nove Mesto Legerova 9 Hotel Louis Leger”, and once you notice that, the set stops being a pile of images and becomes a street-by-street survey with the addresses attached.

Where these numbers come from. All 897 Czech rows come from Wikidata, which catalogues buildings rather than businesses. Each row carries a name, a coordinate pair, a city label, the filename of one image on Wikimedia Commons, and — unusually for a Wikidata country — often a website. There is no address column: every address in this article was read out of an image filename, which is an inference, and we say so wherever we use one. Queried 17 August 2026.

The best-photographed country we hold, and the emptiest

The table is the shape of the whole catalogue in miniature, and it runs the wrong way round. The countries at the top are photographed because they come from Wikidata, which is built by people who upload pictures; the countries at the bottom come from national tourism registers, which collect facts and no pictures at all. France is 1.5% photographed and carries quality labels, opening data and classifications for tens of thousands of properties. Czechia is 99.1% photographed and carries nothing but a name, a point on the map and a picture.

With one exception, and it is worth correcting the record here: 469 of the 897 Czech rows do carry a website, and inside Prague the share is higher still. Czechia is not quite the factless country its picture rate suggests, and it is the only one of our Wikidata countries where a reader has somewhere to go for the real details. Telephone numbers, descriptions and star grades remain at zero across all 897.

“Prague” is not one city in this data — it is fifty-five of them

You cannot count Prague by asking for it. Only 120 rows carry the city value `Prague`; the rest of the capital is filed under its districts and quarters, as though each were a separate town.

So we stopped trusting the label and drew a box instead: everything between 49.94°N and 50.18°N and 14.22°E and 14.71°E, which is Prague and a thin margin of its suburbs. That box holds 456 rows, and all 456 carry a picture — a clean 100% — with 298 of them carrying a website. That is the real Prague figure. Anyone quoting 897 for Prague, as our own topic brief did, is quoting the whole of Czechia: the Giant Mountains ski villages, Karlovy Vary, Plzeň, Brno and the spa town of Janské Lázně are all in that number.

The filenames carry the addresses the database lacks

Not one Czech row has an address column. But 327 of the 889 Czech image filenames begin with the word Praha, and a large share of those follow one rigid convention: city, district, street, house number, and then the name of the hotel. “Praha Nove Mesto Zitna 33 Hotel Archibald City.” “Praha Stare Mesto Rybna 10 Hotel Astoria.” Applying a strict pattern match — district from a fixed list, then a street, then a number — resolves 252 files into a district, a street and a house number, across 165 distinct streets.

• This is an inference, not a field. The filename is a caption written by a photographer, not validated data. It can be stale, misspelled or simply wrong, and nobody has checked it against the land registry.
• Czech buildings carry two numbers — an orientation number and a conscription number — and the filenames mix them. One file in our set reads “Sokolska 422”, which is a conscription number and will not help anyone find the door on a street where the orientation numbers stop in the seventies.
• Diacritics are stripped in the filenames and present in the names. Žitná becomes Zitna, Spálená becomes Spalena. Any matching between the two has to be done on de-accented text, which is exactly how we found the duplicates further down this page.

Twelve hotel buildings stand on one six-lane road

Counting the extracted addresses by street produces one result that stands well clear of the rest. Twelve of Prague's hotel buildings are on Legerova — at numbers 9, 13, 15, 19, 22, 32, 35, 39, 41, 54, 60 and 61. No other street in the city comes close.

Legerova and Sokolská are the two halves of Prague's north–south arterial, the magistrála — six lanes of traffic running past the National Museum and down towards Nusle Bridge. Together they hold nineteen of these buildings. The most hotel-dense addresses in Prague are on its loudest road, and they are there for an obvious reason: the magistrála is where the coaches could stop and where the post-war blocks were big enough to convert. Žitná, third on the list, is the same road's westbound partner. What the filenames have quietly reconstructed is not a picture of romantic Prague at all — it is the arterial hotel belt around the edge of the New Town.

One person took most of these pictures

We read the licence and author of a random sample of 60 of the Prague images from the Commons API. Forty-one of the sixty were taken by the same contributor, a Wikimedia editor who signs as JiriMatejicek — the maker of the photograph at the top of this page, and of most of the Legerova series, several of which are dated to a single day in April 2019. The rest of the sample is scattered across a dozen other names.

That is the honest description of what this dataset is. Czechia is not the best-documented country in our catalogue; it is the country where one person went out with a camera and a list of addresses and worked through it. The 99.1% picture rate is a measure of that person's persistence, not of Czech hospitality, and if you are wondering why Prague is comprehensively covered while a comparable city elsewhere is not, this is the whole answer.

• Fifty-nine of the sixty sampled files require attribution — 44 under CC BY-SA 3.0, 15 under CC BY-SA 4.0. Exactly one was released under CC0 with no conditions.
• Share-alike is a condition, not a suggestion. A CC BY-SA image can be reused commercially, but the photographer must be named and the licence stated wherever it appears. That obligation travels with the picture onto every hotel card on this site.
• The sample is 60 of 456, drawn at random. The concentration is unlikely to be an artefact at that size, but it is a sample, and we have not read the author of every Prague file the way we did for Miami Beach.

897 rows are not 897 buildings

De-accenting every Czech name and counting the distinct results gives 878 names on 897 rows, and rounding the coordinates gives 894 distinct positions. The gap is duplication, and the filenames show exactly how it happens: Hotel Residence Spalena and Hotel Residence Spálená are two rows for one building at Spálená 33, separated only by an acute accent that one editor typed and another did not. Ječná 12 carries three rows — a hotel, a pension and a club — which may be three businesses in one building or one business recorded three times, and our data cannot tell you which. V jirchářích 13 carries two, one of them named simply “house at V jirchářích 13”.

What we don't know

• Whether any of these buildings still operates as a hotel. Wikidata catalogues buildings, including former and demolished ones, and we hold no operating-status field. The 298 websites inside the Prague box are the best signal available and they are not proof; check one before relying on it.
• No prices, no star grades, no room counts, no ratings. Every one of those columns is empty on all 897 Czech rows, as it is on almost our whole catalogue — we hold booking links for 13 properties out of 132,375. Nothing here is bookable through us.
• Every address on this page came out of a filename. No Czech row has an address column. The extraction resolved 252 of 889 files; the other 637 have filenames in a different style, so the street table is built on 28% of the pictures and is a floor, not a census. Legerova may well have more than twelve.
• The district labels are inconsistent and sometimes wrong. Fifty-five different values appear inside one city, mixing quarters, cadastral areas and the city itself, and 56 rows inside the box say only Unknown. Nothing in this article is grouped by that field for exactly that reason.
• We have not been to Prague for this. Everything above is a query result and a filename. The claim that the magistrála is loud is general knowledge about a six-lane urban motorway, not something our data measured — and if it matters to your sleep, check the address on a map rather than trusting us.

Sources

Building records from Wikidata, metadata released under CC0 1.0; the photographs are separately licensed and are not public domain. Image licences and authorship read from the Wikimedia Commons API on 17 August 2026 for a random sample of 60 Prague files. Hero image: Hotel Louis Leger at Legerova 9, Nové Město, photographed 26 April 2019 by JiriMatejicek, CC BY-SA 3.0, via Wikimedia Commons. Counts: 897 Czech rows held, 889 with a picture, 456 inside the Prague coordinate box printed above.

More from the same catalogue

• Miami Beach has 287 hotel buildings on record, and the pictures are postcards — the same Wikidata problem with the opposite answer: an advertising archive from 1945.
• Barcelona's Eixample: 52 hotel buildings in one neighbourhood — another city, another single volunteer, 39 of the 52 pictures taken by one man.
• Where to sleep above 2,000 metres in South Tyrol — what these articles look like when the register carries real facts.
• Every Czech property we hold — the destination page behind these counts.