Commit-Level Dataset on Organizational Participation in 975 Open Source Software Projects

This dataset provides a commit-level view of organizational participation in 975 popular Python open-source software (OSS) projects. It was constructed for research on OSS governance, contribution dynamics, and leadership patterns. The list of projects was derived from two curated Software Heritage archive snapshots — the 3,052 most popular Python projects (2019) and 443 popular Python repositories (August 2024). From this combined pool, 975 GitHub-hosted projects were successfully mined using the PyDriller framework. A key feature of the dataset is its systematic organizational attribution. Contributor emails from commit metadata were processed to: Extract and normalize registered domains (e.g., dev.example.com → example.com). Filter out commits from public email providers (e.g., gmail.com, outlook.com) and bot/service accounts. Map remaining domains to canonical organization names (e.g., fb.com → Meta Platforms, Inc.). Retain only commits confidently classified as firm-likely or academic-likely. The final dataset contains 1,924,792 commit records attributed to 10,033 unique organizations.Key fields include: Project, Contributor Email, Commit Date, Insertions, Deletions, Number of Files Changed, Email Domain, Organization, and Label (specifying firm_likely or academic_likely). This dataset enables analyses of organizational concentration (HHI, Top-1 share), contribution diversity, and leadership persistence in OSS ecosystems. It is released to foster reproducibility and to support further research on collaboration and governance in open software development. Source: Software Heritage archive (2019 & 2024 snapshots)Extraction tool: PyDrillerScope: 975 Python projects (1,925,274 commits, 10,033 organizations)Format: CSV (UTF-8)