Using Twitter to Predict Voting Behavior
Mike Chrzanowski, Daniel Levick · 2012
An increasing amount of research has emerged in the past few years using social media to either predict the outcomes of elections or, in the case of the United States, to classify users as Democrat or Republican. While predictions based on social media are not representative of the voting population, they have been shown to compete with surveys in accuracy and can be done in real time to provide instantaneous feedback to political events (Tweetminster, 2010). Twitter provides an excellent platform for solving political classification problems. The realtime availability of large amounts of data as well as built-in content sorting mechanisms (e.g., hashtags, retweets, followers) are major advantages over using other forms of social media. If these advantages can be leveraged to predict how Twitter users will vote, this information could be used to predict an election outcome or predict the likely impact of real-time events on voting patterns. Problem Statement: Our goal is to train an algorithm that could have predicted an arbitrary Twitter user's vote in the 2012 US Presidential Election, using methods that could be applied to future elections. Recent studies have shown the ability to accurately classify users as Democrat or Republican