2 Comments

Summary:

Twitter is fast becoming the focus group of the 21st century, a status solidified yet again during Sunday night’s Super Bowl. The platform saw 453 times the maximum tweets per second it saw during 2008′s game, and sentiment analysis of tweets might have predicted the upset.

sb tweet analysis

Twitter is fast becoming the focus group of the 21st century, a status solidified yet again during Sunday night’s Super Bowl. The platform saw 453 times the maximum tweets per second it saw during 2008′s game, and sentiment analysis of tweets might have predicted the upset. If you want to know what has webizens, at least, excited, Twitter has to be the place to look.

Here’s how Twitter itself broke down the numbers in a blog post Monday afternoon:

  • There was a total of 13.7 million Super Bowl-related tweets between 6 p.m. and 11 p.m. EST.
  • Last night’s peak was 12,233 tweets per second, up from 4,064 during last year’s game and a mere 27 during 2008′s Super Bowl. The peak came during the game’s thrilling final 3 minutes.
  • Super Bowl watchers also tweeting might be more interested in the games other aspects than is the average sports fan: the top hashtags were all about commercials, and Madonna’s halftime show averaged 8,000 tweets per second with a peak of 10,245 tweets.

IBM released some interesting stats of its own leading up to the game last week. As part of an ongoing project with the USC Annenberg Innovation Lab, Big Blue analyzed fan sentiment across 600,000 tweets to determine which players and teams have the most support. Among its findings:

  • Fans were generally more positive about the New England Patriots players, with wide receiver Wes Welker and quarterback Tom Brady achieving 71 percent and 65 percent positive sentiment, respectively. Their New York Giants counterparts Victor Cruz and Eli Manning achieved positive scores of 62 percent and 69 percent, respectively.
  • The Indianapolis Colts might be wise to reconsider their reported plans to cut Hall 0f Fame quarterback Peyton Manning. Although he wasn’t playing in the Super Bowl, Manning received a positive-sentiment score of 63 percent.

But those results were released on Feb. 2. On Feb. 3, IBM’s analysis showed a role reversal that ultimately mirrored the result of the game: Eli Manning (66 percent) overtook Brady (61 percent) in positive sentiment. Interestingly, Las Vegas oddsmakers favored the Patriots — perhaps leading to the team’s early lead in fan sentiment — but Giants ended up winning the game.

Of course, Twitter has potential outside of measuring interest in the Super Bowl and its players. In October, IBM did a similar analysis around the World Series, and some firms are even using Twitter trends to predict the stock market. On a global scale, some have even analyzed Twitter data to draw correlations between user activity and world-changing events such as the Middle Eastern revolutions and the spread of disease. In the era of big data, Twitter is about as big as it gets when trying to figure out what’s actually happening among the people, as it happens.

  1. Reblogged this on Nick Bolton and commented:
    Phenomenal figures. What was also interesting was that, posts OPA, there was a Super Bowl crackdown with US Federal agents shutting down more than 300 websites
    http://rt.com/usa/news/quiroa-superbowl-copyright-stream-471/

    Share
  2. For Twitter streams, sentiment analysis doesn’t make a lot of sense – for a number of reasons.

    We published our data 1 day before IBM’s, and out “prediction” was right.

    http://blog.itrend.tv/twitter-chatter-on-super-bowl-xlvi-giants-wil

    Blogs are a different story, and according to those, “Giants will win” kind of comments outnumbered Patriots’ almost 2 to 1, even one week before the Super Bowl. So this was actually a very easy prediction to make.

    Share

Comments have been disabled for this post