Wednesday, December 4, 2013
As a long-time laborer in the vineyards of the dismal science, I was pleased recently when someone introduced me as "an optimistic economist." Perhaps in an effort to cement that moniker, and to escape the avalanche of "worst in the nation" tax and regulatory ratings, I was drawn to a report that pronounced Maine the second happiest state in the nation. "Is there something new here?" I wondered. "Or are we just a people who like high taxes and lots of regulation?"
The report -- "The Geography of Happiness" -- was written by a group of researchers from the University of Vermont associated with the felicitously named Vermont Complex Systems Center, Computational Story Lab. Now that sounds like a happy place to work. I'll have to plan a visit before my next visit with the grandchildren.
The study is a wonderful example of the potential of Big Data, of the opportunity to draw meaning from the enormous volume of information made available through our "every breath is memorialized" digital world. The Story Lab researchers gathered, from Twitter's "garden hose feed" -- the mind recoils from the thought of what other data gathering "feeds" may exist -- over 10 million "geotagged tweets" from 373 urban areas in the U.S. This represented "roughly 10 percent of all geotagged tweets posted in 2011." Wow, even with a 140-character limit, that's a lot of words. So what do they mean?
Turns out Amazon has compiled a word "sentiment" (meaning happiness) index for the 10,000 most commonly used words found in the combination of Google books, The New York Times, Twitter and music lyrics. "Rainbow," for example gets a "happiness" rating of 8.1 on a 1-to-9 scale, and "earthquake" gets a score of 1.9. You get the idea.
Having gathered the words and their ratings, the calculation is simple: Tote up the word scores and rank the areas. Turns out that the only people who talk, or, more precisely, tweet, more happily than we songbirds in Maine are those from Hawaii. Makes you feel better than the thought of another snowstorm and a delayed mud season, right?
So, as with all Big Data prestidigitation, the question arises, "What does it mean?"
And, as with all rankings, the first response is to quibble with the data. Not all Mainers tweet, so the conclusions don't apply to everyone. And what about all those out-of-state tourists tweeting about the moose they just saw? Just like with sales tax data, our state score is likely to be affected by visitors.
Then there's the question of word meaning. Louisiana, for instance, "is revealed as the saddest state primarily as a result of an abundance of profanity relative to other states." But does profanity mean unhappiness? If the characters depicted in the movie "Beasts of the Southern Wild" are reflective of Louisiana Twitter accounts, my guess would be no.
Interestingly, Maine has the third most unique word usage pattern of all the states, meaning the pattern of word use third least correlated to the average usage of all 50 states. Does this just mean we (meaning Maine Twitter users) just have a naturally happier vocabulary? Or is that a conclusion we just can't get to from heayah?
The optimistic economist says the strength of "happy" tweets in Maine is further evidence of our quality of place and a reason to be hopeful for our future. The pragmatic economist simply says that this real-time, Big Data analysis business really is the stuff that good stories -- even dreams -- are made on and that our little lives deserve a much closer look.
Charles Lawton is chief economist for Planning Decisions, Inc. He can be reached at: