ValueWalk

Drunk Tweeting? This Algorithm Will Find You Out

As if drunk texting wasn’t bad enough, those with a mind for more public embarrassment sometimes engage in drunk tweeting.

While you might engage in some drunk texting to an ex-partner, or perhaps a potential future ex-partner, if you fancy a spot of drunk tweeting you must be aware that the entire world will be able to see it. In this social media obsessed age, boasting of drunken escapades is not the best way to boost your public profile.

Algorithms trained to spot drunk tweeting

Researchers have now come up with an algorithm that can detect drunk tweeting and guess whether the tweeter was drunk at the time of the post. Nabil Hossain at the University of Rochester, upstate New York, used Twitter and machine learning in order to track alcohol use in a given community.

Hossain and his team collected thousands of tweets from July 2013-July 2014 in New York state. They then honed the selection to those containing alcohol-related terms like “beer keg” or “shitfaced.”

Researchers then passed the 11,000 tweets through three human operatives on Amazon’s Mechanical Turk crowdsourcing platform. They were asked three questions:

Q1: Does the tweet make any reference to drinking alcoholic beverages?

Q2: if so, is the tweet about the tweeter him or herself drinking alcoholic beverages?

Q3: if so, is it likely that the tweet was sent at the time and place the tweeter was drinking alcoholic beverages?

Jumping off point for further study into patterns of alcohol use

Hossain used the answers to train three separate algorithms. They then repeated the process to find out whether the drinkers were drunk tweeting from home or from a bar.

By combining their findings the team were able to determine that people in New York City tend to drink at home, or near home, possibly in one of the thousands of bars. Those that live in the suburbs tend to drink further away from home.

The scientists believe that the technique could be used to find out where alcohol consumption is highest. One limitation is the fact that Twitter users tend to be younger and from some minorities, but Hossain wants to use the technique as a starting point.

“Our future work will perform a comprehensive study of alcohol consumption in social media around features such as user demographics, settings people go to drink-and-tweet,” the paper reads. “We can explore the social network of drinkers to find out how social interactions and peer pressure in social media influence the tendency to reference drinking.”