You may have already seen my coverage of the social media experimentation happening on sites like Facebook and OK Cupid. These large-scale experiments played with their users' emotions to figure out how they would react. Researchers at the University of Michigan, however, just found a non-invasive way to gauge the U.S. unemployment rate through Twitter.
Instead of forcing users to see either "happy" or "sad" content, the researchers searched for terms like "lost my job" and other similar phrases that would indicate someone who had become recently unemployed.
“When we started,” head researcher Matthew Shapiro told the University of Michigan, “we had no idea if we could track job loss with tweets, but over a two-year period, we’ve seen the social media index perform quite well.”
What's interesting is that Shapiro also worked with colleagues to make sure his results were accurate. He added and removed specific keywords like "lost work," which were often typed in reference to a computer crash rather than to talk about losing a job.
Their results were still accurate. While phone and mail surveys have acted as the national thermometer for employment up to this point, but mining Twitter updates can provide an up-to-the-minute report about national employment with surprising accuracy.
I'm glad that these researchers have figured out a non-invasive use to make society a better place through social media. It could save the government and provide everyone with better insight as to how America is doing.