Classified citations with identifiers from English Wikipedia (2024)
Kokash & Colavizza
44.8 million citations from English Wikipedia, tagged as news, book, journal, or other.
Citations extracted from the February 2024 English Wikipedia dump, with templates harmonized and identifiers (DOI, ISBN, PMID, and others) pulled out. A rule-based classifier labels each as news, book, journal, or other. It lets researchers study what Wikipedia rests on.
Access
- Cost
- FreeFree download from Zenodo; no account needed.
- License
- Open licenseCC BY 4.0
- Sign-up
- No sign-up
- How you get it
- Bulk downloadZIP (CSV)
What's in it
- One row is
- One row per citation
- Covers
- English Wikipedia
- Years
- Snapshot February 2024
- Updated
- Annual
- Size
- 44,766,800 citations
Quality
Documented pipeline with yearly snapshots; classification is rule-based and approximate.
- Only identifiers that appear in the citation are captured
- Classification is heuristic
Connects to
- Wikipedia dumps: The raw source this extraction uses
- Crossref REST API metadata: Resolve DOIs for journal citations
- Open Library: Resolve ISBNs to book metadata
Questions it could answer
- What share of Wikipedia citations are to open access papers?
- Which journals are cited most?