In this project, we attempted to build a system which could take any given set of data, plot it on an X axis (usually by date), and then look through a pre-saved datasets for other plots that had similar trendlines. In doing so, it would randomly find datasets that appeared correleated, even though in all likely hood they were not. This project is written in NodeJS, and uses a MySql databse. It uses D3 to draw the data for the client.
In the end though, we did not finish this project at the hackathon, and it currently is an a completely unworkable state - we ruined some things in version control which made things a lot harder right at the end. We might return to this to make it work (we were close), but we also probably will not.