ENGINEERING NEWS

Twitter-based Vote Predictor Named Top Pick by PBS Idea Lab

Algorithm predicts vote based on word choice, hashtags in a user’s Twitter feed

What do your tweets say about your politics? Well, quite a bit.

TweetCast, an algorithm developed by the Knight Lab at Northwestern University, can guess which candidate you voted for in the recent presidential election by using just your Twitter username.

Tweetcast was named on PBS MediaShift Idea Lab’s “Our Picks for the Most Innovative Election Coverage.”

The technology works by examining the words, hashtags, websites, and usernames in a user’s tweets to predict voting behavior. For example, Romney voters are more likely to tweet words like “libs,” terrorist,” and “socialism,” while Obama voters lean toward words like “exactly,” “equal,” and “cuts.”

Established in 2011, the Knight News Innovation Lab is a joint initiative of Northwestern’s McCormick School of Engineering and Medill School of Journalism, Media, Integrated Marketing Communications. The Lab’s mission is to accelerate the transformation of the way news is discovered, analyzed, presented and delivered by introducing smart and practical technology innovation.

Read the PBS Idea Labs article.