Researchers Use Tweets To Predict Unemployment [STUDY]

Researchers Use Tweets To Predict Unemployment [STUDY]

Your tweets reveal more than you think. Every time you tweet about your frustrations at a red light, the cold you just got over or a new pair of shoes you bought, you’re adding to the wealth of data on Twitter. And one group of researchers has tapped into this data to predict just how many people have lost their jobs.

Researchers at the University of Michigan have created the Social Media Job Loss Index, which tracks job losses based on tweets like “I just lost my job! Can you believe it!” and “Well, I am now officially jobless. Time to spruce up the resume.”

The Index plots official unemployment insurance claims numbers from the Department of Labor against a model that predicts unemployment by tracking tweets.

job losses twitter

In the graph above, the red line represents unemployment predictions based on tweets – and as you can see, it’s not too far off from the official blue line.

As the Washington Post observes, the Index uses research compiled between July 2011 and November 2013 that, in part, determined which words and tweets were actually related to job loss. That study uncovered a list of verbal signals, including:

  • Axed
  • Outsourced
  • Lost job
  • Been fired
  • Laid off

The researchers had to filter out tweets that were simply commentary on unemployment (as opposed to those tweeting about experiencing it), as well as slang, misspellings and other factors.

The result is the above Index, which is updated weekly here.

(Source: University of Michigan Social Media Job Loss Index;Need a job image via Shutterstock)