Crossref DOI-to-DOI citations

This dataset contains the citations (references) in Crossref's corpus of the scholarly literature. It is limited to citations between works with Crossref-issued DOIs. It was extracted from a May 18, 2026 snapshot of the full metadata from Crossref's REST API.

The format is a two-column CSV file with header, compressed using gzip. The columns are citing_doi and cited_doi. The DOIs have been lowercased and deduplicated. They can be joined to the full metadata using Crossref's REST API or data snapshots.

The dataset contains 2,000,250,662 unique citation pairs.

• 46% of the cited DOIs were submitted directly by the publisher (doi-asserted-by: "publisher")
• 54% were matched by Crossref using other reference information submitted by the publisher (doi-asserted-by: "crossref")