Those looking for a job can get a good idea of how stressful working for a potential employer may be by looking at its employees’ Twitter traffic.

That’s the implication of a new study published in Applied Psychology by assistant professor Wei Wang, director of UCF’s Computational Psychology Laboratory. His work seeks to make sense of Big Data and how its lessons can be applied to the workplace.

Employers can get a sense of how stressed out their workforce is and adjust if needed. They can also use the information to plan for the announcement of good or bad news. Best of all, the insight can be in real time.

Wang analyzed public data from more than 222 million tweets written by more than 6 million U.S. users in an 18-month period. Tweets were first classified as work-related and not work-related, and then parsed for sentiment and mentions of stress and health problems such as headaches and illness.

“The good thing about Twitter is that it has very accurate time and location information, and it’s highly accessible. It offers a valuable dataset to dig in to,” Wang said.

Among the findings:

  • Tweets talking about work were less likely to mention emotion or stress, but were more likely to bring up health issues.
  • People appear to be least happy during the middle of the week, and express the most positive sentiments during the weekend.
  • Expressions of stress and negative emotions are lowest at the start of the weekend but increase gradually as the weekend progresses.
  • People tend to be more emotionally expressive on weekends than on weekdays.
  • The research team included Wang, Ivan Hernandez of Northwestern University, Daniel A. Newman of the University of Illinois at Urbana-Champaign, Jibo He of Wichita State University, and Jiang Bian of the University of Florida.

    “Right now we are in the discovery phase,” Wang said. “Additional research is needed, but analysis of Twitter data shows promise as a tool for conducting health research.”



    Wei Wang

    Organizations are just beginning to harness the potential of social networks’ Big Data treasure trove. Three years ago, researchers from Johns Hopkins University used sophisticated filtering techniques to analyze tweets about the flu and found their model produced results rivaling data from the Center for Disease Control, but in real time.

    Amit Joshi, a UCF associate professor of marketing, has lectured extensively on social media and Big Data. Social networks such as Twitter offer the appearance of intimate conversation, Joshi said, and that gets people to reveal information about their state of mind.

    “The applications are virtually limitless,” he said. “The data can give you a much deeper understanding of, for example, how good a place is to work for or how satisfied people are with their jobs.”

    Subsequent research may even facilitate the accurate prediction of the unemployment rate or the likelihood that someone will quit their job, Wang added.

    “The future success of businesses largely depends upon their ability to make good decisions using real-time data. ‘Big Data’ has forever changed the landscape of conducting research,” said Paul Jarley, dean of the College of Business Administration.