- November 16, 2011
I’ll admit it – I haven’t been watching the Republican debates. In fact, the closest I’ve come is listening to Auto-tuned versions of Rick Perry gaffs and watching John Stewart criticize Herman Cain’s unfortunate choice of words and conduct around women. Nonetheless, when you look past the political circus playing out before our eyes, you have to remember that underneath it all, one of these people is going to win the Republican nomination and run against Obama in 2012.
140Elect, a Twitter-analysis-based political consulting team, wondered the same thing. To start the journey to an answer, they pulled 7 months of data for each of the candidates and teamed up with Tableau Software to run some statistical analysis and create a data visualization.
Spikes in followers consistently precede shifts in the polls. Look at Gingrich for instance. He receives a major spike in followers and polling when he announces on May 11th, then goes down in the polls until a September 7th GOP debate where he makes a splash. This spike in followers precedes a steady, upward shift in polling for Newt Gingrich.
You can read more of 140Elect’s in-depth analysis here. Using Twitter data to predict the results of elections is not an entirely new idea. In May 2010, UK-based political analysis tool, Tweetminster predicted the Labour party winning the general election that year. When all the results were in, using Twitter data to predict election results was as accurate as traditional opinion polls, with 90.5% accuracy. Later that year in November, social media scientist, Dan Zarrella made similar findings after gathering a random sample of 30 senate, house and governor races. In his analysis, he discovered that in 71% of races, the candidate with the most Twitter followers was ahead in the polls.
Twitter can also be used to predict the voting habits of an individual. TuVoto, an Infochimps-powered site, predicts how an individuals in Spain will vote in municipal elections based on commonly tweeted words and relationships to other Twitter users.
I can see a point in the not too distant future when we start to question the need for and validity of traditional opinion polls when we’ve got the answer sitting in the cloud.