Predicting exactly how waves will break upon the shore is nearly impossible. But forecasting waves that travel along social media streams is more than doable-it's already happening, according to a collection of studies recently published in Internet Research. Consider that virtually every move we make on Twitter, Facebook, and other social media channels is being recorded in some form as documentable data. That data, apparently, can be prescient when collected on a large scale and carefully examined by social and computer scientists who are skilled at examining online behavior. These experts now have the ability, based on identifiable trends, to foresee likely outcomes such as election results, book sales, and movie box office success.
Daniel Gayo-Avello, one of the guest editors for the current issue of Internet Research, says the journal's newly published studies explore various types of forecasting models and ways they are being adapted to social media for the purpose of better comprehending and anticipating human habits and behaviors.
"It is undeniable that social media (provides) an extremely rich source of information, and it makes sense to exploit that data to make predictions or forecast future trends of offline phenomena," says Gayo-Avello. "There are also measurable differences in the way that credible and non-credible (social media) messages spread. Special measures can be used to filter out those messages that are not trustworthy."
Gayo-Avello says the studies also suggest that using statistical methods, in which external (but scarce) data is combined with social media data to build a predictive model, can be a promising approach.
"Up to now, a number of different scenarios-such as (those related to) the economy, politics, health or event detection-have been studied with varied results. For instance, prediction of flu and other epidemic outbreaks seem to be reasonably predictable on the basis of user-generated content," says Gayo-Avello. "Social media can be used to at least predict the present. It can be used as a proxy measure for variables that cannot be measured in real time on a general basis, such as unemployment rates or public opinions on a number of issues."
Additionally, Gayo-Avello notes that other published studies have linked user behavior in social media to attention paid to electronic content. One paper, for instance, shows that web visits and social media comments can be used to foresee future visit to news items. Another study indicates that the number of times a journal article gets tweeted within the first 72 hours after it's published is a fairly reliable forecaster of how frequently it will eventually get cited in other papers.
The findings of the Internet Research papers don't surprise Bill Balderaz, president of Fathom Columbus, a company that is currently using social media to effectively envision outcomes for its clients-including the results of the two most recent presidential elections, which it correctly predicted.
"Social media is the largest, fastest and most uninhibited focus group ever created. (The Internet Research studies) demonstrate that knowing how consumers feel is a predictor of how they will behave," says Balderaz, who's confident his firm can accurately predict everything from stock prices to climate change based on social media conversations.
"It's not as hard as you would think," says Balderaz. "We use historical data correlations to predict the future. We analyze social media patterns going back years, then we correlate social media conversations with real events."
Case in point: Every time social media conversations about stomach aches reached a certain volume, Fathom discovered that two days later there was a flu epidemic. "We can use that (data) to predict a flu outbreak. Or, we know when social media conversation reaches a certain sentiment level, a company's stock price increases," Balderaz says.
Although the findings of these aforementioned studies should excite digital publishers and electronic content providers who may be eager to predict potential sales or consumer proclivities, would-be social media prognosticators need to be aware of the limitations of collected data.
"While it is amazing that we can process the huge volume of data to find these nuggets, it is not like predicting an unknown event like an earthquake," says Michael J. Procopio, a social media strategist. Nevertheless, "we know when books and movies are coming out. And we are now able to ‘see' the buzz and do a better job of matching it to expected results because of the advent of social media networks."
(Image courtesy of Shutterstock.)