What's inside the dataset bundle

Compiled 2026-10-15T06:45:06+00:00 · regenerated automatically · this page is generated from the actual files, so it cannot drift from what you download.

The compiled dataset bundle (single jsonl, indexed sqlite, tidy daily entity csv), refreshed daily — 17,805 rows covering 2026-09-01 → 2026-10-15, one download, refreshed daily.

Buy once — CA$5 → no subscription, no account, no recurring charge

Files in the download

What is in each file

• signals_all.jsonl — every archived signal as one JSON object per line, uniform fields across all sources (schema v1).
• signals.db — the same rows in SQLite, indexed on date, source, org and kind, plus an entities_daily table. Queryable the moment you open it.
• entities_daily.csv — a tidy time series: one row per tracked model or repository per day, with its rank, metric and rank change. This is the shape analysts actually want, and the free per-day files do not carry it.
• MANIFEST.json — row counts, date range, per-source counts, field list.
• README.md — the field list and paste-ready example queries.

Rows per source

Real rows from the file

{
 "date": "2026-09-01",
 "source": "github",
 "kind": "top",
 "title": "codecrafters-io/build-your-own-x",
 "url": "https://github.com/codecrafters-io/build-your-own-x",
 "org": "codecrafters-io"
}

{
 "date": "2026-09-01",
 "source": "github",
 "kind": "top",
 "title": "sindresorhus/awesome",
 "url": "https://github.com/sindresorhus/awesome",
 "org": "sindresorhus"
}

{
 "date": "2026-09-01",
 "source": "github",
 "kind": "top",
 "title": "public-apis/public-apis",
 "url": "https://github.com/public-apis/public-apis",
 "org": "public-apis"
}

{
 "date": "2026-09-01",
 "source": "github",
 "kind": "top",
 "title": "freeCodeCamp/freeCodeCamp",
 "url": "https://github.com/freeCodeCamp/freeCodeCamp",
 "org": "freeCodeCamp"
}

Being straight about what is and is not exclusive

The raw daily files are also free to read (CC BY 4.0) on Hugging Face and GitHub. What you pay for is the compiled form: one de-duplicated JSONL instead of dozens of daily files, an indexed SQLite database, and the daily entity time series — none of which the free mirrors carry. If you would rather assemble that yourself, you can, and that is fine.

Licence: CC BY 4.0 — use it commercially, credit "Signals 4 (Signals API)". Every row links to its original source. Page generated 2026-10-15 06:45 UTC.