Skip to main content

Tracking Happiness Over Time: A Fascinating Twitter Study

Tracking Twitter happiness.
Tracking Twitter happiness.

Social media marketers love to measure things. They love to measure customer engagement, for example. Brand awareness. Sales conversion. Return on investment.

The most fascinating thing to track, however — at least to me — is sentiment. And I’m not talking about that schmoopy, gooey feeling you get when you think about your first crush.

Marketers want to track sentiment is because they’re interested in knowing how people feel about their company or brand. Are they happy or sad? Satisfied or mad? Are they advocating, or are they trash-talking?

Sentiment is difficult to track because of the nature of language — and because of the infiltration of casual words, acronyms, and slang across social media. “Sick,” for example, can be used to describe polar opposite ends of the emotional spectrum. Witness:

“Dude, that was sick!”
[Translation: Friend, what I just witnessed was truly astounding. I am extraordinarily pleased.]

“Dude, that was sick!”
[Translation: You, sir, are a disgusting specimen of the lowest denominator.]

Many tools and analytics systems exist to help parse sentiment data on Twitter, but they tend to work best when tasked with real-time trends such as blockbuster movies, political elections, well-known brands like Coca-Cola and Apple, and high-profile sports events. But there is some disagreement in the marketing world over just how reliable and valuable this analysis really is — after all, Twitter demographics are not wholly representative of the nation’s.

But what if you take just one element of sentiment on social media and track it not just today, at this moment, but over time? takes the model of the usual sentiment tracking tools and turns it inside out. Instead of chasing sentiment about brands, celebrities, or popular culture, the hedonometer focuses on tracking one thing: happiness.

How? The people behind it have identified over 10,000 words with both positive and negative associations and analyzed their Twitter use according to a scoring system. Their slightly more advanced explanation of the system:

“To quantify the happiness of the atoms of language, we merged the 5,000 most frequent words from a collection of four corpora: Google Books, New York Times articles, Music Lyrics, and Twitter messages, resulting in a composite set of roughly 10,000 unique words. Using Amazon’s Mechanical Turk service, we had each of these words scored on a nine point scale of happiness: (1) sad to (9) happy.”

Well, get on with your bad selves.

With the words scored, is able to gauge and plot the happiness averages on Twitter every day since September 10, 2008. The resulting chart is a dizzying array of colorful dots that jump up and down the scale.


Playing with this thing is just plain cool. You can hover over a dot to get details on its score, including the most relevant word shifts. Interested in knowing the happiness average on a certain day? Drag the crosshairs of a section of the slider at the bottom and get a zoomed-in view of your date span.

Unsurprisingly, Friday and Saturday are the happiest days of the week. Tuesday is the unhappiest. Happiness spikes on Christmas Day, Valentine’s Day, Mothers’ Day, Fathers’ Day, Easter, and Thanksgiving. One of the saddest days in Twitter’s history was Thursday, June 25, 2009, when Michael Jackson died.

Interestingly, overall happiness on Twitter has slowly declined over the years. Is that in correlation with changing Twitter user demographics, or is it reflective of the larger society?

This is not your average corporate marketing tool. This will not help you understand how consumers feel about Taco Bell’s Doritos Locos tacos. But if you’re looking for a view of social media from a sociologist’s, historian’s, and data analyst’s perspective, you’re in luck.

The best part is that this is only just the beginning. is only using Twitter right now, but could expand to any data source. There are also plans to add an API soon.

This is one of those tools I’m going to love checking over time. What are your thoughts?


Leave a Reply

Your email address will not be published. Required fields are marked *