# BookRank open data

Snapshot 2026-09-21. 12,048 works (2,538 scored), 1,373 series, 1,154 authors, 49 category rankings.

| File | What |
|---|---|
| books.csv | one row per work: score, ratings, confidence, best rank, identifiers |
| books.ndjson | the same plus titles in every language, categories, all ISBNs |
| identifiers.csv | crosswalk: isbn13 / wikidata / openlibrary / googlebooks / hardcover → work |
| rankings.csv | every rank in every category |
| series.csv, authors.csv | series and authors with their counts |
| datapackage.json | Frictionless Data descriptor |

Score: a Bayesian weighted average of reader ratings on a 0–10 scale, prior weight 184, prior mean per category. A work with fewer than 50 ratings is listed but not scored. Confidence = ratings ÷ (ratings + 184). Method: https://openbookrank.com/en/methodology

Licence: scores, ranks and confidence CC BY 4.0 — credit "BookRank (https://openbookrank.com)". The identifier crosswalk is CC0. Per-source averages are not in these files; they come under their providers' API terms and are shown only as evidence on a work's own record.

API: https://openbookrank.com/api/v1/index.json