Twitter analysis reveals global human moodiness
Baku, October 13 (AZERTAC). An analysis of half a billion tweets shows people wake up happy but get grumpier as the day progresses.
Using the micro-blogging site Twitter as a gauge of global sentiment, US social scientists studied 509 million tweets from 2.4 million users in 84 countries between February 2008 and January 2010, according to research reported in the journal Science.
Twitter, the five-year-old site that lets users communicate 140-character posts, offers an unprecedented chance to study human behaviour and social networks, scientists from Cornell University said.
Online expert Peter Griffin told TV ONE`s Breakfast scientists feed tweets through word monitoring software developed at the University of Auckland and they can tell from language how people are feeling.
Griffin said in the morning people feel good but as they day goes on they get more morose and stressed out and rebound again at midnight.
But before you blame your boss the changes are related to sleeping patterns and how people feel physically.
Those brief posts get down to the nitty-gritty, showing that Twitter users prefer bacon to sausage and Cheerios to Frosted Flakes, Cornell University sociologist and co-author Michael Macy said.
The change in moods was the same across the globe, he said, from India, Africa, Australia and New Zealand to Britain, Canada and the United States.
To get a picture of the global mood, the scientists searched tweets for about 1000 words that reliably indicate positive emotions “agree,” “fantastic,” “super” and negative ones, such as “afraid,” “mad,” and “panic.”
The researchers considered up to 400 tweets from each person, and excluded those with fewer than 25 tweets.
The scientists noted a recent survey of US Twitter users found they are 51% white, 24% African American and 17% Hispanic. People with college and advanced degrees and with higher household incomes show up in higher numbers.
The researchers said this still makes their study more representative than earlier ones on college students.