This is my side project that I have been doing on and off for the last two years.
This is a social networking site somewhat like Twitter/Facebook, but with some applications of Natural Language Processing and Information Extraction and Information Retrieval. Each status messages that are posted to the site, we process the post to mine its meaning and then combine this with the entire pool of user posts in our database to generate some aggregated data based on the user post.
For example, if you write a post saying "I think Obama will win this election", the service will mine its existing data and present you with all related posts from other users. You can see charts of the results based on country, age, and gender (we will add more charting options in future).
Some example posts you can try with - "I love NY", "I hate NY", "I live in NY".
This is still in private beta testing, but I would love to hear opinions and suggestions from you guys! Thanks!