It takes a digital village

An analytical-ish look at what grows strong communities on GitHub.com

Transcript

1. It takes a digital village
   An analytical-ish look at what grows strong communities on GitHub.com
   @benbalter
   ben.balter.com
   ben.balter@github.com
   

   View Slide

2. The promise of open source
   Publish code People contribute
   

   View Slide

3. Online communities are offline
   communities, just online
   

   View Slide

4. All communities have three things
   e.g., neighborhoods, congregations, bowling leagues
   

   View Slide

5. Shared challenge
   a nice place to live, practicing religious beliefs, or an excuse to wear silly shoes
   

   View Slide

6. Leaders
   elected officials, clergy, team captains
   

   View Slide

7. Norms (and the means to enforce them)
   laws, religious teachings, league regulations
   

   View Slide

8. Open source communities
   Challenge - Project goal
   Leaders - Maintainers
   Norms - Meta files
   

   View Slide

9. Encouraging good 

   online citizenship
   

   View Slide

10. The Community Profile API
    

    View Slide

11. Community Profile API
    

    View Slide

12. GET /repositories/:repo_id/community/profile
    [
    {
    "health_percentage": 100,
    "files": {
    "code_of_conduct": {
    "name": "Contributor Covenant",
    "key": "contributor_covenant",
    "url": null,
    "html_url": "https://github.com/octocat/Hello-World/blob/master/CODE_OF_CONDUCT.md"
    },
    "contributing": {
    "url": "https://api.github.com/repos/octocat/Hello-World/contents/CONTRIBUTING",
    "html_url": "https://github.com/octocat/Hello-World/blob/master/CONTRIBUTING"
    },
    "license": {
    "name": "MIT License",
    "key": "mit",
    "url": "https://api.github.com/licenses/mit",
    "html_url": "https://github.com/octocat/Hello-World/blob/master/LICENSE"
    },
    "readme": {
    "url": "https://api.github.com/repos/octocat/Hello-World/contents/README.md",
    "html_url": "https://github.com/octocat/Hello-World/blob/master/README.md"
    }
    },
    "updated_at": "2017-02-28T19:09:29Z"
    }
    ]
    

    View Slide

13. GitHub as a dataset
    21M
    developers
    58M
    Projects
    

    View Slide

14. Three quick qualifiers
    

    View Slide

15. Types of projects on GitHub
    80% - Homework, published code, terrible ideas
    20% - Open source projects
    

    View Slide

16. Only looking at
    Public, non-fork, non-spammy repos
    

    View Slide

17. Correlation != causation
    

    View Slide

18. Opensource.guide
    

    View Slide

19. The contributor funnel
    @MikeMcQuaid
    

    View Slide

20. How to get more users
    

    View Slide

21. Solve a shared problem
    

    View Slide

22. Market your project
    

    View Slide

23. Websites attract PRs
    # of Pull Requests % of repos with homepage
    1-10 23%
    11-100 44%
    101-1000 65%
    1000+ 84%
    

    View Slide

24. Documentation maturity model
    README
    Wiki
    /docs
    

    View Slide

25. How to get more contributors
    

    View Slide

26. Reduce friction
    

    View Slide

27. Friction (n) - the time it takes to go
    from "I want to contribute" to "I have"
    

    View Slide

28. Foster a welcoming community
    Artifacts
    • README
    • Contributing
    • License
    • Code of conduct
    Practices
    • Welcoming
    • Coaching
    • Thanking
    • Mentoring
    

    View Slide

29. Document process
    

    View Slide

30. CONTRIBUTING files attract contributions
    Pull requests Unique contributors Repeat contributors
    1-10 2% 1% 1%
    11-100 8% 16% 22%
    101-1000 25% 42% 34%
    1000+ 58% 6%* 1%*
    Percentage of repositories with a CONTRIBUTING file
    

    View Slide

31. Articulate your vision
    

    View Slide

32. Things to include in your README
    • What problems are we solving?
    • What's the technology stack?
    • What's the project's status?
    • How to use/contribute?
    • If you want contributions
    http://ben.balter.com/2017/04/14/your-projects-readme-is-your-project-constitution/
    

    View Slide

33. Percentage of repositories with a README
    Pull requests Unique contributors Repeat contributors
    1-10 86% 69% 73%
    11-100 92% 86% 84%
    101-1000 97% 85% 81%*
    1000+ 98% 99.75% 99.84%
    

    View Slide

34. License your project
    

    View Slide

35. chosealicense.com
    

    View Slide

36. Percentage of repositories with a LICENSE
    Pull requests Unique contributors Repeat contributors
    1-10 28% 18% 24%
    11-100 50% 64% 66%
    101-1000 67% 75% 78%
    1000+ 82% 97% 97%
    

    View Slide

37. Add a code of conduct
    

    View Slide

38. GitHub Community Guidelines
    

    View Slide

39. Why add a code of conduct?
    • Reasonable people disagree on the internet
    • Establishes expectations for community norms
    • Signals that your project is a welcoming community
    • Better to have before something happens, not after
    • It's the right thing to do
    

    View Slide

40. View Slide

41. Percentage of repositories with a Code of Conduct
    Pull requests Unique contributors Repeat contributors
    1-10 0% 0% 0%
    11-100 1% 1% 1%
    101-1000 2% 3% 3%
    1000+ 8% 0%* 0%*
    

    View Slide

42. How to get more maintainers
    

    View Slide

43. Get repeat contributors
    

    View Slide

44. Repeat contributors are hard to find
    # of single
    contributors
    Code of
    Conduct
    CONTRIBUTING LICENSE README Homepage
    1-10 0% 1% 16% 68% 11%
    11-100 1% 19% 66% 86% 28%
    101-1000 5% 51% 78% 92% 20%
    1000+ 1% 3% 97% 99% 17%
    

    View Slide

45. Decentralize governance
    

    View Slide

46. Large projects are organization owned
    # of pull requests % of org-owned repositories
    1-10 19%
    11-100 50%
    101-1000 75%
    1000+ 94%
    

    View Slide

47. Take aways
    

    View Slide

48. Number of contributors
    Complexity
    

    View Slide

49. Adoption is low
    % of
    repositories
    >= 1 issue >= 1 PR >= 5 PRs >= 100 PRs >= 500 PRs
    Code of
    Conduct
    0% 0% 1% 2% 5%
    CONTRIBUTING 0% 3% 7% 27% 50%
    LICENSE file 4% 31% 47% 68% 80%
    README file 10% 87% 90% 96% 98%
    

    View Slide

50. What's next?
    

    View Slide

51. View Slide

52. Expanded community metrics
    • Issue and PR templates
    • Issues opened by non-collaborator
    • Contributions from non-collaborators
    

    View Slide

53. Product interventions
    

    View Slide

54. Hi there! It looks
    like you're trying
    to start an open
    source project.
    

    View Slide

55. Real-time community insights
    

    View Slide

56. It takes a digital village
    An analytical-ish look at what grows strong communities on GitHub.com
    @benbalter
    ben.balter.com
    ben.balter@github.com
    

    View Slide