Maksim2042 teaches a Social Network Analysis class at GMU, largely based on his book Social Network Analysis for Startups. I’ve been meaning to collect the GitHub data out of curiosity for months, but haven’t found the time. Happily, one of the requirements of his class can be fullfilled by moving my dusty, back burner idea to the front burner.
I am part of a three person group interested in studying the social network properties of coders on GitHub. GitHub has the following relationship types:
- user belongs to organization
- user following user
- user owns repository
- user watching repository
- user forking repository
- user contributing to a repository
From these relationship types, we would like to see if we can predict collaborations. This is somewhat similar to a 2009 Contest held by GitHub. Quoting:
“So, the problem is that we want to recommend repositories to you when you log into GitHub that you’ll love. How do we find the perfect projects for you? I wanted to just look at networks of what people were watching and figure out what you might like by what your friends liked.” — schacon, Build a recommendation system, July 29, 2009.
The main difference is we are interested in generating contributors, not watchers. Contributors advance software, watchers don’t.