Wow, that sounds great! Make sure to share the results with everyone here on HN!
Interesting - how are you doing the sentiment analysis?
1)First few (say 1k or as many one wants) will be done by manually tagging the tweets as +ve, -ve or neutral, through a web or command line interface.
2)Then, I will use Google Prediction API.
3)Based on this data we can get a graph depicting the rise and fall of overall sentiment for a particular stock over a period of time.
1)It finds and prepares database of tweets about public companies by searching tweets containing Company's name or ticker symbol.
2)Tags them as positive, negative or neutral comment.
3)Prepare a report depicting market sentiments about a particular stock over a period of time.
4)Compare it with how stock is actually doing in that time period and after that.
It is an experimental project to check if we can predict how the stock will fare by measuring the market sentiments toward it and if it is reliable or not.