Projects > TweetCast


Obama or Romney? TweetCast predicts which candidate you are most likely to vote for in the 2012 U.S. presidential election based on an analysis of your Twitter activity.

Our technology examines the words, hashtags, websites, and usernames in a user's tweets to predict future voting behavior. To help understand how these features are useful in distinguishing between supporters of the two candidates, we have provided a list of the most discriminating terms used by Twitter users supporting Obama and those supporting Romney.

TweetCast is one demonstration of our broader effort to develop technology that can predict future activity and preferences based on social media analysis. It was created by Shawn O'Banion and Larry Birnbaum at Northwestern University's Knight Lab, with support from the John S. and James L. Knight Foundation and the National Science Foundation. Please send general comments or feedback to

Project Papers

  • Using explicit linguistic expressions of preference in social media to predict voting behavior