A new study ranks cities as happy or sad based on tweets

Zach Lozano

Tweets are usually meant to share random facts and thoughts, but researchers are now using Twitter to geographically gauge happiness. 

A recent study by Lewis Mitchell, a postdoctoral associate with the Computational Story Lab at the University of Vermont, has examined word usage in tweets to determine if people in certain parts of the United States are happier than others.

“This study is a part of a larger effort called hedonometrics where the words people use online in large text sources infer something about happiness,” Mitchell said. 

In order to determine happiness level based on location, Mitchell ran more than 10 million geotagged tweets through an algorithm combined with demographic data from the 2010 Census. Based on the results, researchers were able to estimate happiness levels of different cities in the United States.

“We noticed that cities in the South were sadder than cities in the North,” Mitchell said. “Cities with higher socioeconomic level were happier and cities with a higher level of education were also happier. There was also a relationship with obesity. As obesity increased, happiness decreased.”     

The study ranked Austin as the 21st happiest city in the country based on the high frequency of tweets including words such as “coffee,” “awesome,” “free,” “music” and “sweet.”

Mitchell found that Napa, Calif., was the happiest city in the United States and Beaumont, Texas was the saddest. 

“Napa was happiest because they use less swear words, people used words related to wine and were just overall positive,” Mitchell said. “People in Beaumont used a large number of swear words and a large number of words like ‘stupid,’ ‘bored,’ ‘jail’ and ‘pissed.’”

Robert Quigley, a senior journalism lecturer who focuses on multimedia, said while the study is interesting, it should not be taken too seriously as tweets are a questionable marker of happiness and words that have one connotation can often be taken out of context to have a different meaning that is the opposite of its usual implication.

“Different people use Twitter,” Quigley said. “It has a certain level of demographic that is not representative of a regular population so it is hard to determine if an entire population is happy solely based on Twitter.” 

On the other hand, Andrew Whinston, an information systems professor, said Twitter is a good way to monitor happiness. 

“This way is a better way than sending out surveys,” Whinston said. 

He said tweets are more personal than surveys and are more likely to give people a better depiction of someone’s lifestyle.

Published on February 25, 2013 as "Twitter study offers insight into happiness".