In The Matrix, Keanu Reeve’s Neo uses, in the opening scenes of the film, a search algorithm to track down Morpheus and try to understand the reality of the eponymous matrix. It’s taken just twelve years for reality to catch up.
The idea that social media ‘noise’ and reaction may give rise to predictive trends which are more accurate than anything else we have is not new. Last year, in post on using Twitter to predict Oscar winners, I came pretty close to the actual results, predicting the eventual winner a day before the announcement, and I was not even concentrating very deeply in my search, nor did I have much available time to devote to it at the time.
Today, we surface to a reality where a supercomputer, fed sufficient social media data can predict global unrest, local revolutions and, presumably, market crashes. Kalev Leetaru, from the University of Illinois' Institute for Computing in the Humanities, Arts and Social Science, used a an SGI Altix supercomputer, known as Nautilus, based at the University of Tennessee to which it fed hundreds of millions of articles drawn from the world of news gathering and social media.
The news angle of the articles was required for direction while the social media angle was necessary for what is now called Mood Detection or ‘automated sentiment mining’, basically an analysis of social media articles, responses and noise for words such as ‘terrible’, ‘horrific’ or ‘nice’ which then determine a rise or dip in sentiment in a given social graph in a specific geographic location.
Working retrospectively and going back to data over 30 years Leetaru found that the computer’s analysis could precisely predict the Arab Spring happening in Egypt and Tunisia, it predicted, correctly, the revolution in Libya and even suggested the correct location of Bin Laden within 200 km of where he was found and in direct contradiction of the body of analysis which placed him elsewhere.
Leetaru’s supercomputer uses an algorithm which analyses as many as 100 trillion interconnected relationships around each story, examining the intricate web of social media interaction which springs up around it and the reactions it creates. The average marketer is not expected to do anything like this but that does not mean that social media’s predictive analysis should not be part of the arsenal being used to create better marketing campaigns.
When I accurately predicted the Oscar Night Winner I used just two tools: Tweetreach and Google Trends. Since then we have, as individuals, gained better social media data mining skills which allow far more accurate predictions to be made.
By using information from the US Government’s Open Source Centre and BBC Monitoring in addition to Google Alerts plus TweetReach and Google Trends, even a person working alone in a backroom, has the analytical predictive power of a full blown governmental war room, at his fingertips.
Used right this allows social media marketers to better fine-tune their proposed campaigns to take advantage of rising (or dipping) sentiment, jump on the bandwagon before it even begins rolling, predict trends and suggest media campaigns to clients, understand where the market is heading and how to better position new products for better reception, use sentiment mining to help brands steal a march on their rivals.
In a Social Media Today piece on How Social Media is Changing the World I looked at the impact social media use is having on societies everywhere, eroding traditional barriers which were used to filter the flow of information and acting as a catalyst of change.
Like any tool whose impact is real the very use of social media channels is changing the way we perceive the world and our ability to then measure the change in perception is affecting the impact of social media and its further use.
Whether we like it or not we are locked in a cycle where social media is a constant. What we have to do is use it for all the right purposes.