Making us happy today is a project called HappyStance, recent winner of Hack Day and “a mashup of Twitter geo-location API data and sentiment analysis research,” reports the New York Times. Essentially, it’s an app, created by Jeff Larson, Al Shaw, and Julian Burgess, with the help of Heena Ko and Erik Hinton of the New York Times, that measures the emotional mood of people riding New York City’s subways day and night.
The idea behind HappyStance came from a 2009 Stanford research paper that looked at emoticons to provide “sentiment analysis” using a Bayesian Classifier. Larson thought he’d use that system to classify neighborhood blocks in New York City. Given the timing of Hack Day, Shaw suggested they do it by subway stop.
The team pulled a sample of 6,000 happy and 6,000 sad tweets to classify 100,000 tweets around subway stops. Helpfully, the MTA provided geocoding data for each subway stop. It’s all charted out for your viewing pleasure: green is happy; red is not.
“Basically, it’s artificial intelligence,” Larson told us. As for the happiness/sadness levels they found, he had a few comments:
“Across the board the tweets are happier — when you’re happy about something, you tweet about it more. But the anomalies are fun,” he said. “The subway stops around Port Authority were not so happy. The subway stops around Yankee Stadium were the happiest; maybe there was something going on that Saturday.”
He reiterates: “None of this is ‘scientific’ — it’s just a snapshot of the mood on that particular day, the two times we ran it (morning and night). But you can see a trend; people in Manhattan are a little less happy.” (People also seem to be happier at night, mostly.)
“You use this kind of data for market research,” said Larson. “We thought, let’s do something funny with it. Keeping the irreverence is a big factor in it.”
Check out the app and gauge the mood of your subway line here.