What are the best drunk tweets

Suff tweet algorithmHere put Drunk your Tweets from

Drinking and tweeting - they never went well together. American scientists have now developed an algorithm that recognizes where and when the drunken tweets originate.

"Don't drink and text." And yet it happens again and again. Because the cocktail is so chic, the television program is so stupid, or the ex is so great that it wants to be shared with the world. American scientists at the University of Rochester have now been able to use an algorithm to find out where and when we tweet when we are drunk.

To do this, they first had to subject the tweets to a alcohol test to identify the drunk messages. In a figurative sense, of course, only. In addition to searching for obvious words such as "alcohol", "drinks", "party", "drunk", the crowd also helped with the assessment. Their findings were combined with other parameters.

Per mille test using an algorithm

The users were activated via Amazon's Mechanical Turk tool. With this service, people work on tasks, for example editing texts. Around 11,000 tweets helped identify users in this way. Was it tweeted at home or in a bar? Drunk or sober? The rule was: If three agreed to have a "drunk text" in front of them, the tweet would be included in the scientists' data. Information on this was provided, for example, by the time or the spelling of the tweets.

External content

This leads to external content from a provider such as Twitter, Facebook, Instagram, etc. If you load this content, personal data will be transferred to this platform and possibly other third parties. You can find more information in our privacy policy.

In a second step, geo-tagging was used to evaluate whether people were drinking at home or on the go. Here, too, with the help of keywords: Are you tweeting in front of the telly, on the sofa, in the kitchen or from the bathtub?

New Yorkers drink at home

The result is that New Yorkers drink significantly more (or "drunk-tweet") than the residents of the small town of Rochester, home of the researchers, for example. At first glance, it can be assumed that this is due to the large number of pubs. However, another study result speaks against this: New Yorkers are drinking more at home. The scientists cannot explain why this is in turn.

The underlying data set should also be included in their results: Twitter followers tend to include younger and media-savvy people. And they may have different drinking habits than their grandparents.

External content

This leads to external content from a provider such as Twitter, Facebook, Instagram, etc. If you load this content, personal data will be transferred to this platform and possibly other third parties. You can find more information in our privacy policy.