State of diversity in open source communities
summary
Data from 50 projects (about 30 GB of commits plus 11 GB for geographic data) and roughly $300 of human review uses GitHub extraction, Spark notebooks, name-based gender inference, human labeling and sentiment analysis to report correlations not causation. Positive mailing-list sentiment correlates with higher non-male participation while negative sentiment correlates with lower; inferred non-male contributor share tracks with non-male committers; Jupyter projects show higher diversity than many ASF projects; explicit code of conduct is neutral and mentoring guides unexpectedly correlate negatively.
AI-written summary · about
chapters (27)
- 0:00Intro
- 0:22Disclaimer
- 3:03Metrics
- 4:40The Good Girls Revolt
- 5:40Research
- 7:36Legal advice
- 9:46Human data
- 11:18eyeballing
- 13:37unique
- 14:45science
- 16:39bias
- 17:25first look
- 18:26mailing lists
- 19:27committers
- 20:12inferred data
- 20:29code of conduct
- 21:25programming languages
- 23:55Apache committers
- 25:49Germany
- 26:27Canada
- 26:49Uniques
- 27:32Other facets
- 28:26Links
- 29:51Burn out
- 30:30Book author
- 30:43Teaching distributed computing to children
- 32:16Spark setup
related talks
55mDiversity in OpenSource … show me the data!
25mThis is bigger than us: Building a future for Open Source
07mDiversity in Open Source
38mKeeping Your Open Source Community Accessible to All
23mEquity and inclusion as a way of developing diverse, resilient, and more relevant OSS
32mBuilding inclusive Open Source communities
30mCommit to Inclusion: Building Better Open Source Culture
01h 04mWhat's So Open About Open Source?