Rebuilding a domain Wikipedia cites 8,816 times

By James BurfieldSEO consultant · figures recorded 24 August 2026

There’s a domain that Wikipedia links to 8,816 times. It’d been dark for years.

It’s back, rebuilt to 24,162 pages. Whether those old citations still count for anything is the experiment.

The rebuilt site. Figures for 246 countries, with the source and boundary named on every page.

What it used to be

The domain previously published the World Gazetteer, a reference work that listed populations for places worldwide. Before it went offline it had become one of those quiet sources the rest of the web leans on. Wikipedia editors cited it constantly, and those citations stayed in the articles long after the site stopped answering.

The 2005 edition has been restored alongside the new data, which gives a 20 year view on the same places.

What links to it now

Referring links by source, Ahrefs, 23 August 2026. Plus Wikidata, Wikibooks, Wikimedia, the regional government of Andalusia and the University of Bologna.

Majestic rates it Trust Flow 16 against Citation Flow 13. A ratio above 1 is unusual, and it means the links are mostly the kind you’d want: cited by people writing something, rather than placed by someone building a profile.

What got built

246 countries and 193,820 towns and cities. The Spanish and Portuguese pages read in their own language rather than being translated wrappers on an English page.

The UK section, where coverage is deepest at 4,012 pages.

The problem it actually solves

Ask how many people live in London and there are three correct answers.

Most published lists pick one of those silently, which is why no two of them ever agree. The site’s rule is to give the number, say which boundary it describes, name the source and its vintage, and say plainly where the figure is uncertain.

Sounds like a small thing. It’s the whole reason population data is so hard to quote correctly.

How the sources are ranked

Sources are ordered by how defensible they are, and never mixed inside one table.

Statutory registers come first, because where the law requires registration the count is complete. Food businesses have to register with their council, so the Food Standards Agency register gives an exact count of takeaways rather than an estimate.

Official statistics next. UK annual series come from ONS, NRS and NISRA mid-year estimates through Nomis.

Open geographic databases last, mainly GeoNames for world breadth, with the caveat stated on the page that it draws on national censuses of varying dates.

What is being tested

One question: do legacy citations still carry weight when a domain comes back?

There’s very little public data on this. Plenty of people buy aged domains, but they’ll usually redirect them or build something unrelated, which tells you nothing. Here the new site does the same job the old one did, for the same audience, which is the only version of the test that means anything.

The honest baseline

What would count as a result

Three things, in order.

Whether Google indexes a site of this size when it appears on an old domain, and how quickly. Whether the legacy citations translate into rankings, or sit there as historical artefacts. And whether any of the traffic is worth having, because 193,820 place pages could rank widely and still bring in nobody who cares.

I’ll publish the answer whichever way it goes.