I spent the last year tracking public GitHub activity across 4,200+ startup orgs and backtesting whether engineering acceleration precedes fundraise announcements. The result is a dataset of 219 documented fundraises paired with the GitHub activity metrics observed in the weeks before each announcement.<\/p>\n
What’s in it: for each fundraise event, commit velocity over trailing 14-day windows, distinct contributor counts over 30-day windows, new public repo creation, sector classification, funding stage, and the lead time between signal and announcement (21-47 days in the cases where the pattern fired).<\/p>\n
Dataset (Zenodo, CC BY 4.0): https:\/\/doi.org\/10.5281\/zenodo.19650920<\/a><\/p>\n Methodology preprint (SSRN): https:\/\/doi.org\/10.2139\/ssrn.6606558<\/a><\/p>\n Honest caveats: roughly 23% false-positive rate (acceleration sometimes means an enterprise deal or open-source push rather than a round), stealth companies with no public repos are invisible to this approach, and AI-heavy startups are noisy because model releases create commit spikes.<\/p>\n Everything comes from the public GitHub REST API v3, no scraping or private data. Happy to answer questions about the panel construction or backtest design.<\/p>\n<\/div>\n submitted by \/u\/Worth_Wealth_6811 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
