Every year, Microsoft challenges students to come up with creative technology solutions to the world's toughest problems. In 2009, a friend and I decided to enter the contest, building an entry that helped address the United Nations Millenium Development Goals.
We spent a year building an interactive that mashed up news stories about global issues with sentiment analysis and entity recognition. The app was meant to inspire interest in important problems and serve as a springboard for learning more.
In the end, we won the competition over thousands of international competitors. Really, though, the big novelty check was just icing on the cake. By the end of our work, Kathy and I had learned so much and gotten to share our work so widely that it would have been worth the effort without any recognition.
Building a framework for interactives
I built our app so that it could generate content based on any keyword, so we were able to address all eight Millenium Development Goals, the only team to do so in the contest.
Sentiment analysis and entity recognition of news stories
A major component of the app was a feed of 'heroes' and 'villans': people mentioned in the news with strong negative or positive sentiment. This proved really interesting to explore.
Preempting a Twitter feature
I conceived of and built a sentiment analysis visualization of Twitter before this technique was widely used in marketing tools and even Twitter itself.
I leaned on existing APIs to do much of the heavy lifting for this app.