People as "sensors"

Twitter messages allow users to be used as “live sensors”. Engineers from the University of Rice and Motorola Mobility have found a way to measure the level of anxiety and activity of fans, as well as to track the key moments of the games of the National Football League (NFL) (for example, this is American football ). A program analyzing tweets can tell for a few seconds when a touchdown, interception or other important game moment has occurred, and also show how much fans are excited about this moment.

“It is not customary to think of people as sensors (sensors) that react to the surrounding reality, but thanks to such social media as Twitter, it is now possible to track such reactions from millions of people in real time. That’s what SportSense does, ”says Lin Zhong, co-creator of SportSense, an assistant professor of computing and computer science at Rice University.

In collaboration with engineers from the Betaworks group at Motorola Mobility Research Center, Lin Zhong and his students began developing software for tracking and analyzing tweets in 2010.

“We chose American football because touchdowns, interceptions and other game moments make his fans very excited and make people write a lot of tweets. We found that a thorough analysis of tweets allows us to determine exactly what moment occurred. The program can answer for about 20 seconds exactly what happened. We often find out about this even before this information appears on sports sites, ”- Lin Zhong.
')
Emotions of fans and their preferences for teams are reflected in tweets quite transparently, so the program can even determine in whose favor the event that occurred.

For example, based on an analysis of fan tweets, the Detroit Lions win over the Dallas Cowboys had a very high level of anxiety among other games of the same day. The game between teams had an average of 28 tweets per second during the game, which is 9 times more than average games with low activity.

SportSense will analyze all games in real time throughout the NFL season. All these statistics can be viewed on the website in the section Recaps.
Note: TPS - stands for Tweets per second, tweets per second.

“The research team is interested in using programs for analysis and other events that are happening in the world.
Any events that have a sufficiently large audience, for example, which are covered by television, such as reality shows or political debates, can also be analyzed in a similar way.
We are also interested in tracking events on a local scale. For example, when the light is turned off in my area because of a thunderstorm, I would like to know when it will be restored, even if I am at work at this time. People constantly write about events of this kind and all this information is already in tweets, it remains only to find it, ”- Zhong.

Video in English: