An open, versioned index of public science data — so the record stays findable when the websites move.
Science Almanac is a catalog, not a data warehouse. Each entry is a small, human-reviewed, machine-validated record pointing to an authoritative public dataset — its canonical source, how to access it, where it's archived, and whether it is still reachable today. A daily job re-checks every source and opens an issue when one goes dark.
It covers data from publishers like NSF, NASA science programs, NOAA research archives, and other open science publishers.
catalog.json is the machine-readable index; browse catalog/ for the human-readable source.
The repo boots with a single placeholder entry so validation and CI can run; good first issues
are the fastest way to replace it with real datasets.
You don't need to write code: suggest a dataset with a short form
and a curator will turn it into an entry. Prefer a PR? Adding a dataset is one file in catalog/ —
see CONTRIBUTING.md. CI validates every entry against the schema.
Stewards: see docs/STEWARDING.md.
Catalog, don't host. This repo maps data; it does not store data bytes. See AGENTS.md.
A vertical of The Almanac — a community-maintained commons for public data worth keeping findable. Built from almanac-template.