Using Twitter for Research Reveals Trends in Global Sentiment

These days, any active Twitter user could easily be mistaken to have a serious case of mood swings.

The Internet's social networking sites make it all too easy to make immediate updates on changes in how we feel, whether it's about a particular event, person, or idea. When something good happens and you feel happy about it, you want to tweet it. Same goes for when bad news comes your way-you're inclined to open up your Twitter page and let the world know about it in 140 characters. Well, sociologists think that's a perfect way to study behavioral trends internationally.

Cornell University scientists in New York have just completed a social behavior study focusing on Twitter. Published in the journal Science, the research was based on a two-year period of observations recording 509 million tweets from 2.4 million users. Scientists gathered participants from 84 different countries to determine whether there were any trends or correlations to people's moods across regions.

Researchers coded for positive tweets when they contained words like, agree, fantastic, and super. Posts containing negative language like mad or panic were categorized as negative tweets. They found that the happiest tweets tend to be posted in mornings and around midnight when the night is winding down. The general trend thus goes from a positive post early in the day, followed by a grumpier mid-day tweet, and then ending on a happier note in the evening. One explanation may be that people are in better moods when they away from work.

During work hours, there would be a dip in how cheerful posts were. Of course, the weekend saw more positive messages. The interesting thing to note here, however, is the way in which the same mood swings occurred even on weekend days when there wasn't work. This could mean that mood fluctuations aren't actually caused by spending time at work. This was also observed in several countries besides the U.S., including Africa, India, Australia, New Zealand, and Canada, among others. Seasonal changes showed to have an effect on how positively or negatively tweets were, as well. Nearer to the winter when days shorten, posts were less upbeat.

To go more in-depth with this research, scientists are now planning to focus more on behavior rather than just users' feelings. They are hoping to take the study further by analyzing specific emotions showing signs of anxiety, depression, or anorexia.

Indeed, this study can lead to larger implications beyond just facts about mood swings and social networking. Some researchers believe it can help us better understand the degree to which varying events influence the state of mind. Twitter and social networking technology are scientific tools like we have never seen before. The ability to analyze a mass across regions is not only infinite in scale, but also the makes the overall study generalizeable. For social scientists, this is a definite window of opportunity opening up many exciting possibilities in the future of research.