Status updates, limited as they are in character space, can make up an intriguing barometer for the public’s mood if they’re taken as a whole.
That’s just what Facebook is doing, engaging in some basic sentiment analysis by looking at the share of positive and negative words in status updates across English-speaking, U.S. users. They’ve launched a Gross National Happiness index. (That’s inspired by the gross national happiness metric from the Bhutanese government, which criticized traditional measures of national well-being like gross domestic product.)
Facebook’s program looks for positive words like “happy,” “yay” and “awesome,” and negative, or unhappy words, include “sad,” “doubt” and “tragic.” Some not-too-surprising findings: Its gross national happiness index spikes on holidays like Thanksgiving and Christmas. Barack Obama’s election was more than twice as happy as the average Wednesday.
There are a few independent projects like Twendz and Tweetfeel that are exploring the same concept with Twitter. Two Vermont statisticians are harnessing tweets to create a “hedonimeter” that will be released soon at onehappybird.com. Others are exploring building businesses around sentiment analysis — particularly for companies that want to know how the public thinks about their brand in real-time. Pepsi, in fact, does this by collecting tweets mentioning the brand and tracking sentiment changes week by week.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.