Lead author Peter Dodds, an applied mathematician at University of Vermont, and a team of scientists at the Complex Systems Center gathered more than 46 billion words written in tweets by 63 million Twitter users worldwide for three years.
To get a sense of the emotional gist of various words, the researchers used a service from Amazon called Mechanical Turk, which paid a group of volunteers to rate, from one to nine, their sense of the "happiness" of 10,000 most common words in English. For example, laughter was scored at 8.50, food at 7.44, truck at 5.48, greed at 3.06 and terrorist at 1.30.
The researchers applied the scores to the millions of words from Twitter to show changing patterns of word use that provide insights into the way groups of people are feeling at any given moment.
The study, published in the journal PLoS ONE, found Christmas and Valentine's Day's were the high points of the year, while the weekly high point was the weekend and Monday and Tuesday the low point.
Other low points were the tsunami in Japan in March, the U.S. economic bailout -- the Troubled Asset Relief Program signed into law by President George W. Bush Oct. 3, 2008, and the death of actor Patrick Swayze Sept. 14, 2009.
"All the most negative days are shocks from outside people's routines," Dodds said in a statement. "But the overall happiness signal is on the decline."