This Machine Can Tell If You're Drunk Tweeting Or Just Being Plain Annoying
Thanks to some bold researchers from the University of Rochester, there is now a machine that can tell the difference between drunk tweets and plain old annoying tweets.
According to Quartz, they were able to do this by first gathering geo-located tweets from New York City and Monroe County in New York state.
They then filtered out the tweets that included words like “drunk,” “beer” and “party" -- you know, drunk people vocabulary.
After joining forces with Amazon's Mechanical Turk crowdsourcing service, they analyzed over 11,000 alcohol-related tweets.
That's a lot of drunk-speak to sort through, but what they found was pretty surprising.
Using the algorithm, they were able to differentiate tweets about alcohol from actual drunk tweeting.
They were also able to determine whether people were tweeting from their home or not, which is both technologically advanced and super creepy.
From that information, they deduced more drunk tweeting happens when you're nursing that Franzia at home than at the bar. (I could have told you that, scientists.) They also discovered there were way more drunk tweets going on in NYC than Monroe County.
Ultimately, there is a significant bias in the data, considering Twitter isn't an accurate representation of the entire population and neither is New York City.
Still, this algorithm shows there is much to be learned about the behavioral link between alcohol consumption and social media participation.
Namely, why are we getting drunk and trying to reach out to our exes all the time?! Only time and lots of research will tell.