Fortune tellers once used crystal balls to predict the future.
But modern day soothsayers could use social networks like Facebook and Twitter to forecast events, researchers have claimed.
A team from the Pacific Northwest National Laboratory and the University of Washington have collated the results of hundreds of separate studies and found that social media can “make predictions about the future” – although these forecasts are not always accurate.
“Social media (SM) data provides a vast record of humanity’s everyday thoughts, feelings, and actions at a resolution previously unimaginable,” the academics wrote.
“Because user behavior on SM is a reflection of events in the real world, researchers have realized they can use SM in order to forecast, making predictions about the future.”
Whenever you upload a picture of your cup of coffee to Facebook or share your thoughts on the latest episode of “The Bachelor” on Twitter, you are actually handing over information to “companies, governments, and researchers” hell-bent on gathering “data about people on a massive scale.”
We already know this allows advertisers to understand our wants, needs, and desires in terrifying detail.
But could it allow shady spies and creepy corporations to predict future events?
Sadly, the answer to this appears to be yes.
In a paper that’s just been published on Arxiv, the team of researchers found that social media can be used to “detect and predict offline events.”
Twitter analysis can accurately predict civil unrest, for instance, because people use certain hashtags to discuss issues online before their anger bubbles over into the real world.
The most famous example of this came during the Arab Spring when clear signs of the impending protests and unrest were found on social networks days before people took to the streets.
A system called EMBERS (Early Model Based Event Recognition using Surrogates) has also yielded “impressive results” not just in “detecting events, but in detecting specific properties of those events.”
It has been used to predict unrest in South America, forecasting events with 80 percent accuracy in Brazil and a slightly underwhelming 50 percent in Venezuela.
Another study showed “impressive” results in detecting “civil unrest” linked to the Black Lives Matter group.
Social media has also been shown to predict the weather in certain circumstances, with one experiment using tweets from five British cities to forecast rain.
This basically works by analyzing a year’s worth of tweets and then using this information to guess what the weather will be like in any given month.
“The model does quite well, with the notable exception of predicting the weather in July, which is both a summer month, with individuals tweeting about sunny, outdoor activities, but was also the second most rainy month in the dataset,” the academics wrote.
Social media offers “some value” to police looking to predict future crimes, the researchers continued.
Cops could combine historical data about where and when crimes were committed with social media posts to have a stab at working out the likelihood of offenses occurring at certain places and times.
“Of the 25 crime types examined, 19 see a forecasting improvement by incorporating SM data,” the academics added.
Disease outbreaks are also relatively easy to predict because when you spot a few people grumbling about the norovirus or another infectious disease, it’s a fair bet that a lot more people are about to catch it.
Mental health outcomes in an individual can be predicted by analyzing their posts to Facebook, Twitter or Instagram as well.
Depression, for instance, was identified with up to 80 percent accuracy
“Depressed users on SM can be characterized by decreased social activity as well as increased negative sentiment and use of personal pronouns,” the academics added.
Sadly, in some cases, the old crystal ball was about as accurate as social media analysis.
The movements of the financial markets have proved stubbornly difficult to forecast, for instance.
Researchers attempting to guess the outcome of the World Cup were also left disappointed.
“An attempt to predict match outcomes utilizing Twitter data failed to perform better than random chance for early tournament matches,” the team wrote.
They also highlighted “major pitfalls which have made SM prediction,” but ended their study with a note of confidence.
“We find that SM data has been used to make accurate forecasts across all of the disciplines examined.
“Additionally, topics that can be shown to be directly relevant to SM users and how they interact with SM make more successful predictions, such as user location, user demographics, and civil unrest.”