Image courtesy of West McGowan.
Yesterday, Twitter introduced the Twitter Political Index or Twindex. According to Twitter, the Index is "a daily measurement of Twitter users’ feelings towards the [GOP and Democratic] candidates as expressed in nearly two million Tweets each week."
The Index takes all of the tweets created by users in the last three days, measures sentiment and then examines tweets mentioning Obama or Romney, rating the sentiment of those tweets against the overall sentiment of all tweets published to Twitter in that same timeframe.
The Index then produces a score for each candidate. So, if Barack Obama received a 38, that means 38% of the tweets mentioning him were more positive than all the other tweets recorded during that time frame. Make sense?
The Index received wide coverage, with some seeing this as a key example of the potential of big data. Others see if as merely another data point in the political horse race game. For its part, Twitter claims the Index has been quite close to Gallup polling and while imperfect, is an accurate measurement of overall sentiment towards the two candidates.
We've already talked about the research that concludes that Twitter is not that great at predicting electoral outcomes. I also feel that with such a small percentage of voters on Twitter, such an Index could only be at best a representation of a subset of the population.
But for you data junkies, the Twitter Political Index will provide some interesting insight into the voting proclivities of Twitter users.