Download websites that use Ckan

About Ckan

Open data portals are one of the few corners of the web where the choice of software is close to public policy. A ministry that publishes its spending has to put the files somewhere searchable, with metadata, licences, an API and a URL that survives a change of government, and building that from scratch is a project no department wants to own. Ckan is the answer a lot of them reached for, which is why the entry sits under CMS even though what it manages is datasets rather than pages.

The institutional profile follows from the use case. Ckan is maintained by Open Knowledge Foundation under AGPL and dates to 2006, and the combination of a strong copyleft licence with a foundation steward is what a public body wants: no vendor to be locked into, no licence fee to defend in a budget hearing, and an assurance that anyone offering it as a hosted service has to publish their changes. Deployments cluster around governments, statistics agencies, universities, research infrastructures and the contractors who run portals on their behalf.

Presence also implies an obligation rather than an ambition. A department does not install a data catalogue because it fancies one, it installs it because a transparency directive, a funding condition or a research mandate requires the outputs to be published and citable. Portals of this kind tend to survive redesigns of everything around them, and the domain carrying one is usually stable, institutional and worth watching over years rather than months.

Detection is easier here than for a typical Python application, because the software announces itself. Portals expose their release through the same API that serves the catalogue, which is why this entry carries version data that many content platforms in our index cannot supply. As of October 2026 we see Ckan on 12 sites, spread across 9 distinct releases, with 2.9.2 accounting for a large share of them.

The spread is the interesting part. A wide range of releases running in production at once is what a slow procurement cycle looks like from outside: a portal gets built by a contractor, the contract ends, the software keeps running, and an upgrade happens when a new project funds one. For anyone selling managed hosting, upgrade work or data platform tooling into the public sector, the version column is a qualifying question already answered. For a security researcher, an old release on a government domain is a finding by itself.

What the aggregate answers is different in kind from a lookup. Checking one portal tells you what that ministry runs. The list tells you how many public bodies publish open data through a common platform at all, how the population divides between national portals and municipal ones, which countries are represented once the domains are grouped, and how quickly the installed base moves between releases when exports are compared over time. Nobody publishes that census, because no single institution owns the question.

The limits are ordinary but real. Portals behind authentication, on internal networks or rebranded past recognition will not appear, and heavy theming can hide the fingerprint entirely, so treat the figure as a floor and the version mix as a sample rather than a full survey.