We came up with this idea a little late in the hackathon; it was the progression of an idea involving the Jawbone UP and a user's emotions. The idea was that you could make people happy if they were sad or excited if they were bored - you'd find out their mood by analyzing their social media. What we've built is the first step towards this goal; we have created an app which predicts if someone's in need of a little cheering up. Target users would be members of our generation, people who say off the cuff things on social media and who could use a little stress relief.
We didn't get as far as we wanted due to our inexperience with data science, the time constraints (and time consuming nature of munging data) along with some trouble we had setting up our server. Still, we managed to get a data model that works, even if we don't fully understand it, and our web app successfully integrates with Facebook's Graph API while pulling in data from Twitter to help feed our predictive model.