The sum total of digital links, text, data, images and everything else you touch online everyday may just be the closest thing we ever get to a crystal ball.
Yes, there is a serious move afoot – led by very serious people – to mine the Web in order to predict future events. Don’t believe me? Just check out this paper by Microsoft Research and the Israel Institute of Technology.
Their idea: Software can recognize undetectable patterns in the mass of real-time and archived news and other information online – and analyze it in an unbiased way – to forecast future events. H/T to tech blog GigaOM for the link, as well as a great post about how the paper’s authors will be digging into two decades of New York Times articles and seeing what kind of forecasts they can extrapolate.
This idea isn’t so crazy – witness last week’s post about Google’s uneven efforts to predict the spread of the flu – and there are a number of startups in this area, including Recorded Future, which is already being marketed to the defense and financial sectors. You can even try a very limited version of Recorded Future for yourself – although the corporate plans and are going to cost you, big time.
But I encourage anybody to give the demo a spin to get a taste of predictive analytics. The focus of the research is on large, worldwide events, but we’re already seeing plenty of predictive analytics in the business world, including behavioral targeting and even algorithms that attempt to predict what will trend on Twitter.
Just imagine how hard it’s going to be to stay ahead of the curve when everyone knows where the curve is going.