First lady Melania Trump, President Trump, Vice President Pence and Karen Pence observe a moment of silence on the South Lawn of the White House on Oct. 2, in memory of the Las Vegas victims. (Chip Somodevilla/Getty Images)

There are certain days of the year when Twitter is generally pretty happy. Christmas. Mother’s Day. Thanksgiving. The measurement on that comes from the Hedonometer, a tool created by researchers from the University of Vermont and the Mitre Corporation. In short, the team asked people to evaluate the emotional weight of certain words and then, on a daily basis, calculates how common those words are on Twitter. More positive words means a happy day, so Christmas, with all of the tweets including the words “merry” and “happy” and “holidays,” ranks pretty high.

Christmas also gets a boost from the word “Christmas” itself, the 32nd-happiest word of the Hedonometer’s 10,222-word vocabulary. The happiest word is “laughter”; “happy” is fourth after “love” and “happiness.”

The least happy word is “terrorist,” which is preceded by such words as “murder,” “death” and “killed.” So, in the wake of the massacre in Las Vegas on Sunday night in which a gunman killed 58 people, the Hedonometer measured its all-time lowest index for the happiness of Twitter. Before Monday, the day with the least happiness on the social media network was the previous record-holder for most people killed in a mass shooting in recent U.S. history: the attack at the Pulse nightclub in Orlando in 2016. Two of the others in the top five were terrorist attacks: the assault in Charlottesville and the killings of police officers in Dallas last year.

The fifth was the day after the election of Donald Trump as president.


The past 18 months on Twitter. Click to view a larger version.

Peter Dodds is one of the researchers who created the Hedonometer. He spoke with The Washington Post by phone to explain how the tool worked and what he’d learned from years of watching it watch us.

Dodds noted repeatedly that the point of the Hedonometer was less to track the mood of the world than to develop a system for gauging society’s well-being. The tool doesn’t incorporate every tweet, only a sampling of about one-tenth of a day’s tweets. It’s a subset of Twitter, which a subset of the population itself. “Twitter’s not everything,” he said. “It’s some glimpse into people’s thinking.”

That said, it still reveals something about the public mood. “It comports well with Gallup polls, well-being surveys,” Dodds said, but with a different methodology. “We’re not asking people how they’re doing, we’re looking at what they’re saying.”

On Monday, Las Vegas dominated that conversation. The tool shows which words registered the biggest increases in use: “shooting,” “gun,” “victims,” “terrorist” (as people debated the intent of the attack. Interestingly, “Vegas” is a generally positive word in the Hedonometer’s estimation; had the attack happened somewhere else, it would probably have registered even more sadness than it did having happened there.)

Dodds noted that other generally positive words spiked in the “really awful context” of the killings, including “prayers” and “families.”

“It is a very sad time, it’s an awful time — but there’s a lot of goodness being expressed in terms of people’s responses,” he said.

Twitter’s mood recovered fairly quickly after the Las Vegas shooting. By Wednesday, it was back to where it had been Sunday. That’s common after a tragedy, Dodds said. But it didn’t happen as quickly after Charlottesville.

Dodds said it was “very clear” that this was in part because of Trump’s response. Twitter’s happiness had started to head back up, but when Trump a few days later backtracked on his condemnation of racists to say that “many fine people” had demonstrated alongside the neo-Nazis in those protests, Twitter’s happiness went back down.


The broad trend in the Hedonometer looks a bit like a sine wave. Higher from 2009 to 2011, a slow fade to 2013, a rise into 2015 and then back down. The early years, Dodds said, were volatile as a function of the shifting demographics of Twitter as more people joined. He attributed the recent slide, beginning in early 2016, to the sheer volume of events on which people were offering their thoughts.

Beyond holidays, I asked, what days were the happiest that the Hedonometer had measured? Dodds pointed to several: the wedding of Prince William and Kate Middleton, for example, or the day same-sex marriage was legalized in the country. Also, things like the birthdays of celebrities — in particular, boy-band members. In 2015, the second-saddest day marked by the Hedonometer were the terrorist attacks in Paris that killed more than 100 people. The saddest was when Zayn Malik left One Direction.

Dodds was quick to note that Twitter was always a mishmash of various subjects and interests and feelings. He pointed out that several days before the Charlottesville unrest, there was a downturn in Twitter’s mood, in which people were talking about “war” and “nuclear” (again, driven by Trump tweets).

“There’s always stuff going on everyday,” Dodds said. “Nuclear war’s being talked about, but someone’s birthday is being talked about as well.”

Certainly a fair summary of modern life.