Community Data is Not Community Metrics
summary
Claims that raw dumps of GitHub/GitLab data become useful only when you ask precise questions and run repeatable tests; Red Hat uses Grimoire Labs data and Project CHAOSS concepts to build atomic “Lego brick” metrics that combine into stories about evolution, diversity, value and risk. Examples include mailing list traffic jumping from ~1,000 to 1,500 then 2,000 messages, treating downloads as noise versus tracking release cadence slipping from 6 months to 8–9 months, and flagging a bus factor of 1 as real risk.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:45Blue Dragon
- 1:23Mantis Shrimp
- 2:33Data Breaches
- 4:01Community Data
- 5:40Questions are critical
- 7:18What makes communities healthy
- 8:50Evolution
- 12:08Diversity
- 15:48Combining Metrics
- 19:10Questions
related talks
36mCommunity Data: What to Measure and Why?
43mCommunity Health Reporting 101
35mRegulations, Ethics And Messy Data: Tales From An Open Source Analyst
40mDiscovering Emerging Open Source Communities Through...
01h 04mWhat can we learn from 750 billion GitHub events and 42 TB of code
01h 03mWhat can we learn from 750 billion GitHub events and 42 TB of code
36mPreventing Random Acts of Metrics
41mWhat Can Open Source Project Health Metrics Reveal About Project Users?