visualizedata.org

Sources

Every number here comes from a public publisher. Most arrive through the datasites APIs, a family of free, keyless APIs that each re-read one kind of public data, keep every version, and record the publisher's own files by checksum. This site reads 21 of their 29 datasets. Its checks then go back to the publisher's files and recompute every figure without the API.

The data APIs

economicdata.dev

API: api.economicdata.dev/v1/catalog · 5 datasets, 3 used here

carbondata.dev

API: api.carbondata.dev/v1/catalog · 2 datasets, 2 used here

climatedata.dev

API: api.climatedata.dev/v1/catalog · 6 datasets, 6 used here

airdata.dev

API: api.airdata.dev/v1/catalog · 2 datasets, 1 used here

electiondata.dev

API: api.electiondata.dev/v1/catalog · 3 datasets, 3 used here

satellitedata.dev

API: api.satellitedata.dev/v1/catalog · 3 datasets, 2 used here

waterdata.dev

API: api.waterdata.dev/v1/catalog · 3 datasets, 3 used here

wildfiredata.dev

API: api.wildfiredata.dev/v1/catalog · 4 datasets, 1 used here

visualizedata.dev

API: api.visualizedata.dev/v1/catalog · 1 dataset, 0 used here

Read directly from the publisher

What the checks read

The site's checks never trust the API a number came through. For each publisher they fetch the original files below, accept them only if their checksum is the one the API version recorded (or, for a file the publisher has revised since, say so and limit what may differ), and recompute every figure from them.

Catalogs read 2026-10-27. A dataset listed but not yet published is in its site's catalog with no version yet: it is shown here so the gap is visible, and used once it is published.