alyssafrazee / github_analysis

code for analysis of github repo ownership and gender

gender of GitHub repo owners

---

This repository contains the code and data used to analyze the gender breakdown of owners of public GitHub repositories. I wrote a blog post about what I found.

To reproduce the analysis, run scripts in the following order:

• get_github_info_byday.py: uses the GitHub API to scrape repository data. (nb: This will take something like 60 hours to run).
• merge_files.sh: puts all scraped data together in a big text file
• make_database.R: dumps the scraped data into a SQLite database
• analyze_data.py: processes the data
• bargraph.js: JavaScript/D3 code used to make the graphic showing the results. Alex Wilson made major contributions to this code.

data

The data I scraped in get_github_info_byday.py and processed with merge_files.sh and make_database.R is available in a .db file here. I removed all repo owner last names.

dependencies

Python libraries: PyGithub, Unidecode, Pandas, SexMachine, Matplotlib. Make sure these are installed before running scripts. See requirements.txt for a more detailed specification of Python dependencies, including versions.

R packages: devtools, proto, DBI, chron, RSQLite, and RSQLITE.extfuns. All can be installed from CRAN in R using the install.packages function.

github_analysis's Issues

Yo. IDEA.

What if we changed this to be checkboxes? And then it would control sorting by the sum of the elements rather than each one individually. Because you could recreate this situation, but also do so much more!