Twitter Tea Leaves Put the 49ers and Ravens in the Super Bowl

A social media analytics company looked at what people are saying on Twitter about the New England Patriots, the Baltimore Ravens, the San Francisco 4ers and the Atlanta Falcons &mdash and used to predict who'll win the playoffs and the big game.
Image may contain Airport Airfield Transportation Vehicle Aircraft and Airplane
The ball flies just beyond the grasp of San Francisco 49ers tight end Delanie Walker (46) during the third quarter against the Arizona Cardinals on Dec. 30. The data crunchers at HootSuite say the Twitter tea leaves show the Niners will win the Super Bowl.Photo: Marcio Jose Sanchez/Associated Press

The data crunchers at HootSuite have read the Twitter tea leaves to predict the future — or at least the outcome of the NFL playoffs. They're betting the San Francisco 49ers will win the NFC this weekend. The Baltimore Ravens will win the AFC. And when the two teams meet in Super Bowl XLVII, the Niners will prevail.

Twitter isn't psychic, exactly, but data taken from conversations may have predictive qualities, which is one reason the Pentagon wants to analyze it. HootSuite, a social media analytics company, examined Twitter conversations about the four playoff teams — the New England Patriots, the Baltimore Ravens, the San Francisco 4ers and the Atlanta Falcons — and made its prediction based on what was being said, and how often, about them.

More on Stats in Sports
Basketball Isn't a Sport. It's a Statistical Network
Soccer Embraces Big Data To Quantify The Beautiful Game
Analytics Reveal 13 New Basketball Positions
Stats Prove Jordan's Olympic Team Beats Kobe's
Post-Interception 'Momentum' Is a Myth"Twitter is giving us an outlet now to be able to amplify our thoughts and opinions," said Sandy Pell, public relations manager at HootSuite. "You can project your voice so much further and have people react and engage."

So just how do you analyze Twitter sentiment? HootSuite focused on eight emotions that filled fans' chatter: affection and friendliness, enjoyment and elation, amusement and excitement, contentment and gratitude, sadness and grief, anger and loathing, fear and uneasiness and (my favorite, as a sports fan) humiliation and shame.

HootSuite chose these points of measurement with the help of Lymbix, an analytics company that pays attention to feeling and tone based on positive and negative expressions in Twitter posts, adapting to detect social nuance when language changes. Lymbix wouldn't reveal its secret sauce for choosing these benchmarks, but we have their pie graphs for a more detailed look.

For example, although the Falcons were the topic of much conversation after beating the Seattle Seahawks in the last eight seconds of the game, half of what people were saying was negative. On the other hand, those who weighed in on the Niners kept their posts relatively positive. It was the same story with the Pats and Ravens. Last weekend, there were no tweets expressing humiliation or shame from Niners or Patriots fans.

Grief and anger, however, were the determining factors in prediction who'd make it to the Super Bowl. For the Patriots, the grief and anger no doubt arose from tight end (and star scorer) Rob Gronkowski's season-ending broken forearm. From this we can surmise the folks at HootSuite are betting the loss of Gronkowski will contribute to the fall of the Pats this weekend. Falcons fans undoubtedly were incredulous when the team, up by one point with 15 seconds left, made an onside kick after giving up a 21-point lead in the fourth quarter. Fans of both teams probably are a little less confident in their squad going into the weekend, and that's reflected on Twitter.

These predictions are based on keyword searches, not scientific analysis. The graphs were made after Sunday's games, and undoubtedly would be different if they were generated today, as people look forward to the excitement of the coming weekend.

"We're going to continue playing with the data and prediction," said Pell, "like with elections or the Olympics. There are so many different things you can watch happening on Twitter. It's a lot of fun and a great way to break down the data."

So are they right? We'll have to wait and see.