If you do pretty much anything online these days, you'll be prompted to share it with your social network of friends, family, creepy eStalkers, and the NSA. It's no surprise then that somebody is taking this treasure trove of free information we're spewing all over the Internet and using it to discern some interesting, surprising, but mostly terrifying things. Stuff like ...
Carsten Koall/Getty Images News/Getty Images
By cross-referencing the deluge of posts on various social media outlets against local news stories, surreptitiously intercepted telephone calls, and whatever swear words you use in front of your Kinect, workers at the CIA's Open Source Center in Virginia are able to assess the mood and relevance of virtually any place on Earth.
In other words, using social media, the CIA can identify whether social unrest is apparent in the whole of the community, or just the result of vocalization from a fringe minority with little genuine support for their cause.
Like moderators on the world's crappy message board.
In other other words, by reading our stupid Facebook updates, the feds are able to tell if the whole population is about to revolt, or if somebody just had a really shitty Hot Pocket and thought the Internet should know about it.
Spoiler: They're all shitty.
For example: While the OSC couldn't identify exactly when an uprising would occur in Egypt, Doug Naquin, the center's director, did successfully use the system to predict that "social media in places like Egypt could be a game-changer and a threat to the regime." Sound familiar? That was back in 2011. Surely you remember how that turned out: Something big happened and you turned your avatar green for two days, before replacing it with, like, a really cute selfie.
"Sorry, 'Arab Spring,' but I've got 'Kony 2012' to slacktivise."
A recent paper from Cornell University found that psychopaths have a unique, discernible writing and speaking pattern:
"I found the journal; you need help."
"Sure, take the blender's side."
No, but seriously, after analyzing the speech and texts of 14 psychopathic and 38 non-psychopathic murderers, it was found that psychopaths are more likely to use filler words like "um," are more likely to use cause-and-effect words like "because," and are less likely to make mention of religion or family. Because uh ... um ... oh, right: It signifies their emotional detachment from society or whatever.
"Of course I avoid bringing up God. No one likes it when someone talks about themselves."
The study was then applied to Twitter: a bottomless online pit of rage, narcissism, and delusion. Surprise! They found psychopaths.
Over 3,000 Twitter users allowed their tweets to be examined via a personality test that classified their tweets with three traits: psychopathy, narcissism, and Machiavellianism. As an added incentive (like anybody needs more reward than "find out if you're a Twitter psycho!"), one lucky participant would be given an iPad, because sane or not, everyone needs their Candy Crush fix.
Based purely on their tweets, 41 users were classified as legitimate psychopaths. The real shocker here isn't that, according to reason and logic, some people on Twitter be crazy -- we could have clicked on pretty much any Trending Topic and told you that. The shocker was that most of those 41 results were later confirmed as possible psychopaths by more rigorous and reputable follow-up testing. You know what that means? Just glancing at somebody's Twitter feed and saying "sounds like a psycho" seems to be an actual viable method of diagnosis.
"It's almost as good of a filter as 'nice guy.'"
We can shorten that process even further, to one simple step: Do they follow Kanye West? Yes? Boom. Psychopath.
Jupiterimages/Polka Dot/Getty Images
In a study of a half-million social media users over two years of data, researchers were able to predict how the unemployment trend in the U.S. and Ireland would rise or fall -- just by analyzing how much people complained on their respective social media channels.
It was a lot of data.
Researchers analyzed tweets and status updates where people discussed income or career, and assigned each tweet a "mood score." So, for example, if scientists see a tweet like this ...
... they would assign that tweet a label like "anxious" or "lucky it was only one." (Not all science is hard, folks.) But by using this simple method, researchers found that they could predict something as dramatic as an increase in the unemployment rate. Sounds a little abstract, doesn't it? You could look out your window right now and predict an unemployment rate hike based on the number of hobos fighting over your recycling. Eventually, you would be proven right. But here's the crazy part: The system predicted that a rate hike would come to pass in four months. Pretty narrow window, right?
It happened in three.
And it wasn't a fluke: The findings held true in both the U.S. and Ireland.
Hemera Technologies/AbleStock.com/Getty Images
"Great job proving the Twitter theory. Unfortunately, due to budget cuts ..."
Now, to be clear, these weren't just people specifically talking about their work or bank accounts. Anything income, budget, or job related in the slightest was counted and quantified, and using nothing more than the information we accidentally spill out for fun on the Internet, scientists were able to calculate when we would lose our jobs to within a financial quarter.
Luckily, the same theory holds true in reverse: Supposedly when positive social media posts pop up, they could theoretically predict drops in the unemployment rate.
It's inversely related to the number of ironic scarves and horn-rimmed glasses.
The stock market has always taken Twitter way more seriously than it should. If you're frantically selling off a billion dollars' worth of interest in a company because @cheesypete thinks Velveeta is bullshit, you deserve all the mockery we can give.
"You sold our Apple shares at a loss!?!"
"I had to. @XBox420 called the iPhone 5S 'totes ghey.'"
But actually, there might be something to it: Social scientist Johan Bollen was able to measure the mood states of Twitter users through an algorithm called, shockingly, the Google-Profile of Mood States. GPOMS effectively analyzed 9.7 million tweets by 2.7 million Twitter users over the course of 10 months and established a fairly reliable baseline of "Twitter Mood." Some of what they found was pretty intuitive stuff, like users being panicky and anxious before Election Day or generally happy and calm before Thanksgiving. Really, we're anxious when there are massive changes in the leadership of our country, and content when we get to spend an extended weekend drunk and hip-deep in a gravy pool? Thanks, Science!
"Hop in, Nana's getting the biscuits."
But when the researchers decided to compare the national online emotional state to the Dow Jones industrial average, they found something surprising: The Twitter Mood "calmness" synced with the fluctuation of the stock market -- but three or four days in advance. To clarify, after measuring the nation's overall mood level via stupid goddamn tweets, the algorithm could predict with 86.7 percent accuracy if the market would close up or down several days before it happened. It was, effectively, seeing the financial future by reading the misspelled racist tea leaves of Twitter.
It's even more impressive if you consider that all of this went down in 2008, a period when the market was massively unpredictable -- what with the Afghanistans and the complete economic meltdowns and all. So, rather than focusing on stuff like "historical precedent" or "keen business savvy," it turns out the best economic solution is listening to the misspelled advice people type into their phones while pooping.
And you wonder why nobody respects you, business majors.
Justin Sullivan/Getty Images News/Getty Images
Google Flu Trends is an algorithm that tracks online complaints of flu-like symptoms by monitoring the use of related search terms (e.g., "headache," "temperature," etc.) and plotting them against location. Logically, where search rates are higher, there's a much greater chance of a viral outbreak. That's right: We're using random Google searches -- the same ones that autocomplete "why you" to "why you hurt foot when poke at with stick" -- to successfully track and predict the spread of disease.
The foot's number three enemy, behind LEGOs and garbage water.
However, there are limits to the technology. Researchers saw a spike in search trends when Rihanna tweeted that she had the flu, which, unless it was a dangerous new strain named for the Rihanna species -- like bird or swine -- is less than helpful information to science. So for refinement, analysts turned once again to Twitter. It turns out the 140-character limit is perfect for tracking symptoms: It's longer than the average search term, allowing users to relate the keyword to themselves. In other words, instead of "headache" or "too much poop," you get "I have had a splitting headache all week" or "I have just pooped for the 18th time today and it is terrible." Yet Twitter's character limit also forces users to remain relatively concise with the information they are sharing, so there's no elaborate, wasteful sonnets to the frankly egregious amount of feces exploding out of them at any given time.
The Intelligent Systems Laboratory at the University of Bristol used some 50 million geo-located tweets to create a database of health-related communications. They then compared the database to regional National Health Service statistics over the same period to identify keywords seen during outbreaks of flu, and used those terms to analyze new Twitter activity to predict where and when an outbreak may occur.
In short, if everybody in Michigan is tweeting some form of the word "poo-fountain," scientists know to send extra medical-grade underpants to Lansing.
"But, sir, what about Detroit?"
"Don't you get it? There is no more Detroit."
Paying really close attention to Twitter may not sound as esteemed and dignified as we like our science, but it's much quicker than the traditional method, which is driving through at-risk neighborhoods and holding your hand out the window to see if anybody sneezes on it. A U.S.-based study showed that an outbreak of H1N1 (swine flu) during the 2009 epidemic could have been identified several weeks earlier than traditional methods if we'd just listened to Twitter users.
No word yet on whether or not science considers it "worth it."
"So you can do everything right at your desk-"
Slightly less shockingly: A similar study tracked tweets about vaccinations available for H1N1, and found that regions that showed a negative attitude toward the jab were most likely to be the same regions that had the highest instance of the virus. This is known as a "no shit" scenario, and is commonly attributed to the works and studies of Dr. Sherlock.
Dennis Fulton has given in to peer pressure and now has a meth addiction and a Twitter account where he posts his theories on why his teeth keep falling out.
For more odd indicators, check out 6 Bizarre Factors That Predict Every Presidential Election and 7 Bizarre Trends That Predict an Economic Collapse.
If you're pressed for time and just looking for a quick fix, then check out 4 Oddly Insulting Recent Ad Campaigns .
And stop by LinkSTORM and how Facebook also predicts when you will kick the bucket.
Do you have an idea in mind that would make a great article? Then sign up RIGHT NOW and pitch your first article today! Do you possess expert skills in image creation and manipulation? Mediocre? Even rudimentary? Are you frightened by MS Paint and simply have a funny idea? You can create an infographic and you could be on the front page of Cracked.com tomorrow!
Related Reading: Interested in how social media is changing the world of prison? This article is filled with pictures of convicts bragging on Facebook. Still not convinced that social media is crippling society? Click here and see the proof that REAL friends and online friends are inversely related. Ready for some fiction after all that 'reality' bullcrap? Gladstone's Notes from the Internet Apocalypse will show you the terror of a world without social media.