Researchers: Twitter Can Be Used To Predict Crime


microblogging and crime

Twitter posts can be used to predict crime, says researchers. They aren’t talking about #420 posts either.

It isn’t as simple as looking for tweets where people talk about committing crimes. Outside of some incredibly stupid exceptions, people rarely tweet about crimes they plan to commit. Instead, researchers at the University of Virginia demonstrated that geo-tagged tweets, when combined with data on criminal activity, can help predict when and where crimes are likely to take place before they happen.

Analyzing Twitter was effective in predicting 19 out of the 25 different types of criminal activity tested in the study. It isn’t surprising that alcohol related crimes are among the most easily predicted crimes. Have a large amount of people in one place tweeting about being drunk, and it stands to reason that more alcohol related crimes would take place there.

Stalking, theft and “certain kinds of assault” were some of the other crimes effectively tracked by Twitter analysis according to Business Insider.

The study looked at geo-tagged tweets and crime data from the Chicago police department,  The researchers said they hope police will someday use their algorithms to better allocate police resources. They also said that similar techniques are already being used in Iraq and Afghanistan by the United States military, in order to identify hotspots of trouble.

Minority Report, this is not. However, it is an interesting use of social media data and public funds.

[Photo Credit: eldh]


Ian DeMartino

Ian DeMartino is a Technology, Political and Sports Junkie who only wishes he had more time to devote to each subject. When Ian isn't saving puppies or brokering peace deals in the Middle East, he can usually be found tinkering with electronics or playing video games. Check out his blog at https://techippie.net/

0 Comments

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.