A Dutch professor has discovered it may be possible to predict the ups and downs of the stock market based on what Twitter users are saying.
The “Twitter Predictor”, Johan Bollen, a Dutch Informatics professor at Indiana University in the U.S, discovered “by pure chance” a significant method to predict the U.S. stock market. Meanwhile the American media, including Bloomberg and the Washington Post are closely following his research.
Bollen uses existing psychological research to group words found in tweets – almost ten million of them from around the globe – into six categories which describe how the user is feeling.
In collaboration with Twitter, the professor together with a team of students analyzed data from nearly three million Twitter users over ten months in 2008 and it turned out that they had a prediction accuracy of 87.6% with regards to the up-and-down movement of the Dow Jones four days out.
Previous studies have found a link between Twitter traffic and a movie’s success or failure at the box office but the Indiana University study is believed to be the first to study Twitter and Wall Street.
Bollen is surprised by all the attention. He gets many phone calls from people who see business opportunities in his model.