Some day we won’t have to wait until the monthly unemployment report to know what’s happening in the job market.
New research shows cyber chatter can predict the rise and fall of unemployment in real time.
Many of us complain about our financial health in social media, including everything from cutting back on healthcare to deciding not to buy that new car. Such discussions can provide a window into employment conditions months before traditional government statistics.
Those are the findings of a recent study by software firm SAS and a United Nation's data initiative UN Global Pulse that looked at online discussions in the United States and Ireland over a two-year period and found increased social media talk about things such as postponing vacations or taking more mass transit predicted spikes in the unemployment rates in both countries.
It’s all about giving governments, social service groups and policy makers economic information while its happening so they’re not responding to drops in employment months after they occur, said Anoush Rima Tatevossian, a spokesman for the UN Global Pulse, a UN Secretary-General program to mine digital data in an effort to do good.
“We don’t have enough real time indicators of what’s happening to people as it’s happening,” he said, adding that the study’s findings are evidence digital data is a key source for finding out about the economic here and now.
The idea is similar to what Google has done with its flu tracking initiative, known as Google Flu Trends, that tracks searches related to flu terms in real time in an effort to pinpoint rates of flu nationally.
In the SAS-UN study, researchers looked at half a million entries on web forums and blogs over a two-year period to pinpoint fluctuations in the jobless rate, said I-Sah Hsieh, global manager, international development at SAS.
With this type of information, he continued, policy makers will be able to know right away if the steps they’re taking to deal with economic crises are working and they’ll be able to figure out how much time there is to implement new policies. “With analytics we can quantify how much of a runway we have,” he explained.
In the United States, for example, an increase in angry chatter about workers’ jobs preceded a spike in unemployment by three months, he maintained.
And there are certain leading and lagging indicators of a jobless uptick.
The researchers found an increase in the amount of hostile and depressed chatter in the United States predicted an unemployment spike four months out. On the other hand, talk about housing loss or car repossession happened two and three months, respectively, after a rise in the jobless rate.
In the future, such research will include popular social networking sites such as Twitter and Facebook, said Hsieh. “We’re already starting talks on that, particularly Twitter.”
While he admitted this type of data mining might sound like Big Brother to some people, he stressed that researchers are just analyzing information that’s already in the public forum.
For the UN, it’s a chance to get in on the big data bandwagon and use the information to do deal with global economic issues, added Tatevossian. “Everything is moving faster because the world is so connected now,” he pointed out. “Economic shocks can appear in one part of the world and reverberate all over.”
Real-time data will uncover growing joblessness more quickly, he added, "and we'll be able to intervene and protect people faster.”