USC using sentiment analysis to measure baseball
We’ve seen Twitter used to predict election outcomes, but can Twitter (whose openness and modularity of content makes it the most promising for these studies) be used to determine who will win the World Series?
That’s a question that USC professor Jonathan Taplin and his students are looking to answer using new IBM social media tools.
Rather than looking at raw volume of mentions for a given team, Taplin is using machine-based sentiment analysis to determine whether the nature of the tweets are positive or negative. Whilst this technique is a poor determinant of the outcome of a game (phew, good to know that the outcome of a game is as much a function of the quality of play rather than the social media effusiveness of the fans), sentiment is a predictor of the TV ratings. The team is also testing the relative popularity of players and seeing how closely this matches ‘pro
To see more on BA Forum, the premier IBM business analytics event, check #baforum on Twitter.