Our system has been processing business news articles for several months now, identifying the companies being reported on, as well as the nature of the news--good or bad. One of the first interesting things to extract from our data is "news volume breakouts", when the news volume for a company on a given day exceeds its 50 day running average volume by one or more standard deviations.
We have created a @marketoutsider twitter user to broadcast this information. The tweet broadcasts follow the @stocktwits convention of prefixing the ticker symbol with a dollar sign. Additionally, the ticker symbols are followed by either (+), for an overall sentiment on today's news for the company is good, or (-), when the news is bad. More than one plus or minus indicates a more positive (or negative) sentiment.
For example, today's broadcast was, "$IBM (++) $YHOO (+)", meaning that both IBM and Yahoo had higher than average news volume, and that the overall news for IBM was very positive, while the news for Yahoo was moderately positive.
You may follow the news volume breakout broadcasts at: http://twitter.com/marketoutsider
