Can Social Media Sharpen Quant Models?

'The Sum of All Tweets'

But the notion of traders using sentiment analysis to weed out trading opportunities is a concept that's undoubtedly gaining momentum. And social media sources are a component of that, though probably not the basis for an entire trading strategy.

Rich Brown, the head of Elektron Analytics at Thomson Reuters, says his firm, and companies such as Deutsche Bank and JPMorgan, are exploring the use of unstructured data from sources including Twitter, the news and blogs as a slice of an overall quant model.

"It means that it's one piece of the recipe, and in that case it's much more about the interaction of this signal with the signals that everybody else has in their quant models," Brown says. "Is the run-up in pricing the result of a bunch of news coverage or social media coverage, and where do you see that information flow going over time? These are all much more powerful ways of exploiting that kind of context and commentary than a simple macro that's just judging the sum of all tweets."

But in order to effectively deploy social media in the way that Brown suggests, sentiment needs to be measured on a massive scale, even globally, Progress Software's Bates notes. Since a few individuals could spout bogus ideas and facts, firms might need to measure what a million people might be saying in order to obtain an accurate picture, he explains.

"With news, you're getting it from a Dow Jones — assuming it's over a secure channel, you know that's from a trustworthy party," Bates says. "In social media like Twitter, I can tweet that Facebook stock is going through the roof, and I just made it up. It's not a trusted source."

As the Senior Editor of Advanced Trading, Justin Grant plays a key role in steering the magazine's coverage of the latest issues affecting the buy-side trading community. Since joining Advanced Trading in 2010, Grant's news analysis has touched on everything from the latest ... View Full Bio

2 of 2