Sometimes a picture is worth a thousand words.

This interactive visualization of the location of contributors to the top repositories on GitHub was my entry to the GitHub Data Challenge

The provided Makefile has everything needed to fetch data from GitHub Archive, load it into a database, geocode locations and output the resulting JSON. It can even push the results to GitHub pages.

There are also some interesting queries for doing things like fetching top 200 repositories, gathering up all the relevant events or getting the unique set of locations to geocode.

# This first step takes hours and downloads ~6.5GB
make githubarchive

# This step takes a few hours with decently fast disks
make loaddb

# This step geocodes any locations not already geocoded (`locations.json`)
make geocode

# Generate the `events.json` file used for the visualization
make jsonify

• Geocoding by Nominatim, © OpenStreetMap contributors.

• Inspired by earlier work by Jens Finnäs and Nanda Yadav

• Built with d3 by Mike Bostock