Replication Package: Team Experience Composition and the Stability of Collaborative Knowledge Production

Replication package for a study of how team experience composition relates to the stability of collaborative knowledge production. It contains the full analysis pipeline (pure standard-library Python 3) and the article-level aggregate datasets used in the paper, including a curated English-Wikipedia sample (458 articles) and a Simple-English human-team census (62,979 articles), together with within-article panels, LLM communication-climate labels, and an AI-era monthly series. All data are aggregate and non-personal, computed from public Wikimedia dumps and the MediaWiki/REST APIs; no usernames or individual-level records are included. Code is licensed MIT; the derived datasets are licensed CC BY-SA 4.0 (inherited from Wikimedia). See README.md and DATA_DICTIONARY.md inside the archive.