Social Media Unleashes Social Scientists

twitter

By Carol Boissiere

twitterIn a socially driven world, what do your tweets really say about you?

The Advocate in Baton Rouge, LA recently posted an interesting article on how Twitter posts can be used to track moods:

If you asked a group of Social Scientists from Cornell University, they will tell you that after two years of extensive research and analyzing close to 2.5 millions tweets, they have concluded that individuals wake up happy, but their moods often deteriorates as the day moves forward. A text analysis program was brought in to sort keywords based on emotional content. Keywords such as “happy,” “excited,” “fabulous,” and “smitten” have positive vibes, as well as words like “afraid,” “doomed,” “remorse,” “angry,” and “fury” that have negative vibes.

Patterns certainly emerged and were built around the sun! Tweets with positive overtones peaked twice a day, worldwide, one in the early morning and again around midnight. These same positive peaks came on the weekends, but later in the morning. These patterns showed that sleep schedules and day to day rhythms were important influences on mood, regardless of daily stresses. They also found that moods improved based on daylight savings time. Moods improved in the spring, and worsened in the fall.

Millions of tweeters are part of a study conducted by a group of “Social Scientists” to track moods. Group leader Scott Golder’s original intentions were to study behavior, not emotion. Partnered up with a fellow graduate student, they wrote a computer program that sampled up to 400 messages from Twitter accounts created in 2008 and 2009. This program pulled information such as keywords to see what people were doing and they used time stamps to analyze activities to particular times of day, as well as location tweets from around the world. Did you know bacon was more popular than sausage? They even estimated that it takes approximately 7 hours to become intoxicated based on the time lag between keywords “beer” and “drunk.”

Whether we need or are interested in social scientists telling us when we are in a better mood during the day, this type of research will open up interesting results in the future.