Gitsift

Visualizing GitHub repositories with D3, Crossfilter and Fuse

Gitsift is a visualization of over 119K GitHub repositories using D3, Crossfilter and Fuse. We admit it, code and design is blatantly copied from Mike Bostock's excellent Crossfilter's example. We were interested in leveraging this visualization on GitHub's repositories dataset, and to provide few additional features that help "to go through (something) very carefully in order to find something useful or valuable" (definition of "sift" from Merriam-Webster). You can also look at it as a showoff of how GitHub's Explore page could look like.

Click and drag on any of the charts to select a subset of repositories satisfying certain criteria. List on the bottom shows 40 of these selected repositories that were created most recently, but you can also sort by other criteria by clicking on one of the categories in the header. Also, you can search the url's of these selected repositories to obtain 14 best matches (although bear in mind that it can be quite slow if performed on the whole dataset). The whole dataset is around 3.5 MB so it can take some time for the page to load.

Feel free to play and explore! For example, you can try to find a time period when Javascript surprassed Ruby as the most popular language for new repositories. Or you can wonder why were so many of the repositories back in the 2008 created between 6 and 7 PM?

Time of Day
Repository size
Forks
Issues
Watchers
Created at
Most popular languages
Search selected repositories

Creation date

Repository name

forks

size

issues

watchers

language

Fork me on GitHub