Hedge fund uses Twitter to profitably predict market swings

Story from PRI’s The World. Listen to audio for full report.

A hedge fund is making lots of money using a new program that mines Twitter to determine emotions among the population to help decide buying and selling strategies. The $40 million dollar fund established by Derwent Capital made a profit of almost two percent in its first month of trading in July — that’s way better than how the average hedge fund performed that month.
 
Johan Bollen is the computer scientist at Indiana University who came up with the program.

“It works by scanning millions of Tweets on a daily basis — at least the content of those tweets — and looking for indications of the author’s mood states. From combining all of those together, we determine gauge the public’s mood state on that particular day,” Bollen explained on The World.

While there are key terms the program looks for, the Tweets don’t have to be about financial news, says Bollen, “it could also be ‘I really hate Justin Beeber’s new album'” — which would be indicative of a negative or positive mood of the general public.

The public’s mood state has a predictive correlation with regard to the Dow, according to Bollen.

—————————————————————-

PRI’s “The World” is a one-hour, weekday radio news magazine offering a mix of news, features, interviews, and music from around the globe. “The World” is a co-production of the BBC World Service, PRI and WGBH Boston. More about The World.