Wikipedia Talk Pages Corpus
- Type
- corpus
- Venue
- Cornell University / ConvoKit
- Year
- 2026
- Source
- convokit
- Access
- free
- Language
- English
- Added
- 2026-07-17T20:18:03.685726+00:00
- Verified
- 2026-07-17T20:18:03.685726+00:00
Summary
The Wikipedia Talk Pages Corpus contains conversations from Wikipedia editor talk pages with rich metadata including admin status and edit counts. It comprises 391,294 utterances across 125,292 conversations involving 38,462 speakers (Wikipedia editors). Distributed alongside the paper 'Echoes of Power: Language Effects and Power Differences in Social Interaction' (WWW 2012), it is used for studying conversational dynamics, power differences, and linguistic style coordination in collaborative online environments.
Keywords
wikipedia talk-pages conversations social-nlp power-dynamics convoKit
Topics
NLP / Conversation Analysis
Research notes
- Available via ConvoKit (corpus download: 'wiki-corpus'). A legacy version is available at cs.cornell.edu. Includes SpaCy dependency parses as processed fields.