You don’t have to think about it as an embarrassing situation. According to researchers, your drunk tweets could help scientists determine the ways in which people drink and socialise, and also help them avoid dangerous situations.
A researcher from the University of Rochester, took Twitter data and combined it with machine learning, in order to develop an adaptive algorithm that can define the level of your sobriety and your location when you’re tweeting. The algorithm already contains more than 11,000 geotagged tweets across the state of New York, parsed down to include alcohol-related words and phrases.
If you like our stories, there is an easy way to stay updated:
The algorithm looks for three distinct factors that helps scientists analyse the behaviour of drinking. First of all, it looks for any references to alcoholic beverages, then determines whether the tweets refers to the author drinking or someone else drinking, and then looks at the probability of the tweet being sent at the time and place where the author was drinking.
Once combined, these factors can reveal the ways in which New Yorkers drink (at home, with friends etc.), as well as their favourite places to do so. Going a step further, this algorithm could help scientists reveal the activities that go hand in hand with alcohol consumption, and help you stay safe.