Regulations, Ethics And Messy Data: Tales From An Open Source Analyst
Talkat KubeCon + CloudNativeCon North America 2021 · on Thursday, 2021-10-14· at 11:00· 411 Theater + Online·
summary
Shows concrete pitfalls and fixes for building open source metrics: Git history and GitHub APIs change over time, historical snapshots disagree, and bot noise skews counts — an estimated over 120 million bots generated activity in 2020 and bot accounts cause 44% of push events that year. Recommends tooling and patterns like DevStats + GitHub Archive with Postgres and Grafana, tokenization, compartmentalized access, anonymization, and treating public contribution data under GDPR-like controls.
AI-written summary · about
chapters (11)
- 0:00Start
- 4:35Contribution focus will determine your data sources
- 5:47Project parameters are dynamic
- 8:49Risk classifications are dependent on the parties involved
- 13:29Collection mechanism will impact information volunteered
- 20:03Git activity logs are not consistent and change over time
- 22:40Publishing metrics: you get what you incentivize Measurement and communication methods can influence the behavior you are measuring
- 24:38Who is accountable for open source data?
- 25:47Layers of choice, policy and regulation
- 27:53Read the fine print
- 28:13Good (Data) Practices
related talks
24mA Beginners Guide to Open Source Metrics and Analysis
30mData Tales from an Open Source Research Team
35mThe Benefits and Pitfalls of OSS Project Metrics: Measuring Health and Risk in Open Source Communities
27mA Guide to Responsible Data Collection In Open Source
27mCommunity Data is Not Community Metrics
08mAll eyes on Open Source: Monitoring community sovereignty
26mBuilding & Supporting Open Source Communities Through Metrics
43mDon't Start Over, Start Left: Choosing the Right Open Source...