Twitter is probably one of the primary sources by which people source and share news stories on the web and indeed personal information too. The ability to predict crime may sound like something straight out of the 2002 sci-fi movie, Minority Report, but it may be science fact. New research conducted by the University of Virginia claim that tweets from the popular social media website Twitter, can be used to predict crime. The full report can be accessed here.

Twitter

Last month, the results of the research was published in the scientific journal Decision Support System and according to Techspot.com, it says: “that geo-tagged tweets, combined with historical crime data for an area, can be useful in predicting 19 to 25 kinds of crimes, including stalking, thefts, and certain kinds of assault.” 

Although no one is actually tweeting about carrying out some kind of crime, lead researcher Matthew Gerber says that crimes can be predicted indirectly through monitoring the behavioural patterns of users. “If I tweet about getting drunk tonight, and a lot of people are talking about getting drunk, we know there are certain crimes associated with those things that produce crimes. It’s indirect”.

Researchers analysed the tweets from the city of Chicago and successfully predicted the areas were crimes were most likely to take place. The computer search algorithms correlate the twitter activity of users in a given area and from there, they can make well informed predictions. The research was funded by the US Army who employ similar methods to predict where the next likely threats will occur in war zone situations, such as those in the Middle East.

The Twitter research is expected to be taken further to New York City, Gerber will attempt to recreate the success they achieved with the project in Chicago.

Please leave comments with your thoughts on this story.

Share

Leave a Reply

Your email address will not be published. Required fields are marked *

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