Analyze Foursquare statistics and determine where and when a user is likely to checkin next.
creep.it was an early data science project that I started back in 2011.
The idea was to collect your recent checkin history, throw your checkins into a time vector, plot it as a circle in Cartesian coordinates, run k-means clustering on it, then, using the current time and your location, tell you where it thinks you are going to check in next and when.
But I got busy with things and it fell to the wayside. I may revisit this one in the future.