Predictive Analytics: Using Twitter Data (Part 1)

Extracted 15NOV2011 from http://online.wsj.com/article/SB10001424052970204138204576598942105167646.htm...

Though the practice is still experimental, Twitter data already have become a key variable in behavioral finance investment formulas. "The hedge funds are leading the way," said Chris Moody, chief operating officer at Gnip. Mr. Moody declined to name Gnip's financial customers. "They don't want anyone to know their secret sauce," he said...

Researchers at Indiana University and the University of Manchester who developed the fund's technique say that they can reliably predict changes in the stock market by up to four days, based on the ups and downs of the national mood as expressed through key words in texts sent by 130,000 regular Twitter users...

Other researchers remain skeptical of Twitter's purported predictive power...

A relatively small group of 20,000 users commands the most attention, researchers at Yahoo Research have discovered. They are neither the most prolific nor the most widely followed users, but the website links they recommend are more often repeated and shared by others. When it comes to focusing public attention, content matters more than celebrity, the studies suggest.