Millions of tweets reveal global mood trends

Researchers at Cornell analyze millions of tweets in 84 countries and find that seasonal variations in day length and sleep cycles may affect humans rather uniformly across cultures and borders.

Elizabeth Armstrong Moore
Elizabeth Armstrong Moore is based in Portland, Oregon, and has written for Wired, The Christian Science Monitor, and public radio. Her semi-obscure hobbies include climbing, billiards, board games that take up a lot of space, and piano.
Elizabeth Armstrong Moore
2 min read

It may not be terribly surprising that many of us find our moods dipping over the course of the day, and that by nightfall we light up again. Or that our moods are perkiest on weekends, regardless of which days our weekends fall on (i.e., Fridays and Saturdays in the United Arab Emirates).

Scott Golder Lindsay France/Cornell University

What's of note, according to an analysis of 2.4 million tweets in 84 countries by researchers out of Cornell, is that these mood trends hold steady across cultures and borders, hinting at some sort of deeper trend whose basis is in being human, not in belonging to a particular people or place.

"We saw the influence of something that's biological- or sleep-based; regardless of the day of the week, the shape of the mood rhythm is the same," Scott Golder, a doctoral student of sociology, said in a news release. "The difference between weekdays and weekends has to do with the average mood, which is higher on the weekends than the weekdays. Even in the face of different social and cultural demands, the results are consistent across days."

Twitter opened its data doors to the researchers via a public interface, and researchers processed the incredible volume of data at the Cornell Center for Advanced Computing's Web Lab.

While other tweet-based mood studies have used small, homogenous groups of U.S. undergrads, resulting in inconclusive data, this one was so large, with observations made on an hourly basis, that the team is reporting on its findings in the journal Science with high confidence.

"We've never had the opportunity to measure behavior in this way before," Golder says. "Digital traces of online activity let us do social science in a new way, and to ask questions we've always wanted to ask. How do entire societies work? How are relationships patterned? We're starting to get the data to answer these questions."

Beyond investigating mood trends at certain times of the day or on certain days of the week, the team also looked at whether daylight influences mood patterns. (They are, after all, in Ithaca, N.Y.) What they found is that the amount of daylight doesn't influence mood as much as whether the days are getting shorter or longer. So nine hours of daylight in the spring may produce different moods than nine hours of daylight in the fall.

The researchers have also created the Web site timeu.se to allow users to enter keywords, such as "traffic," "sex," and "fishing," to analyze how behaviors, in addition to moods, are distributed throughout the day.