An assistant professor at Southeastern Louisiana University recently discovered that he could track the flu with relative accuracy on Twitter. Using a bunch of formulas and logarithms we dont understand, Professor Aron Culotta analyzed 500 million tweets and a few keywords to determine how much of the general population had the flu.Culotta noted the whole process, which produced a 95 percent correlation when compared to statistics from the Center for Disease Control, was both cheap and easy. He and his team of two assistants ran a software program that analyzed that days tweets and calculated how many of those users likely had the flu.Aside from monitoring the word flu, Culotta also cross-referenced tweets that contained headache, fever, cough or sore throat. He then used a different strategy to weed out posts that mentioned these words, but actually had very little to do with whether the tweeter had the flu.Culottas next goal is to track location-specific tweets to determine regional trends in flu outbreaks. Click here to read Culotta”s formal findings.
Via CNETPhoto courtesy of Flickr user 7son75