The California News Service plans to use Twitter to help them not only figure out who will win this year, but determine the actual vote share. The "Voting With Your Tweet" experiment is the brainchild of UC Berkeley doctoral candidate Mark Huberty. He analyzed Twitter from 365 Congressional races in 2010 and found that his model picked the winner (after the fact) 92% of the time.
From Pacific Standard:
Huberty “trained two machine learning algorithms to determine what word features of those tweets best predicted whether the Democratic or Republican party candidate won each race.” So in its current incarnation …
"We think the finished algorithm works like this:
First, it identifies from the language in a candidate’s tweets whether they are the incumbent or challenger. Since incumbents win about 85% of the time, this provides a good baseline.
It then adjusts the baseline prediction based on sentiment and action-related phrases. For instance, “voted hcr” (indicating that the incumbent voted for health care reform) was one of the most influential predictors alongside incumbency-related phrases. The algorithm weights those phrases positively or negatively, depending on how predictive they were of a candidate winning or losing."
Right now, the model predicts all four Republicans will win in Utah this year. The vote share prediction (as of October 31) says Rob Bishop will beat Donna McAleer by a 58-42% margin. Chris Stewart is predicted to top Jay Seegmiller 56-43%. Jason Chaffetz will run away with his race gathering 88% of the vote. The site picks Mia Love to beat Jim Matheson 52-48%.