Is Twitter a good enough social sensor for sports TV?

The globalization of TV programming (especially in the Sports and Reality segments) is leading to a bewildering amount of choice for TV watchers. The social currency around knowing what´s happening in these programs as they happen combined with notable undulations in the interestingness of these programs leads to a navigation problem for the TV watcher. The pervasive use of Twitter in conjunction with TV watching makes it a potential sensor for real-time TV, and therefore a building block in solving the user problem of interstitially navigating to the most interesting program at any point in time. Given that users have to tune into shows with peak moments as soon as those peak moments happen, interstitial navigation has to be high enough performance to enable TV event detection and user tune in within tens of seconds of the event. The question being addressed in this paper is - Do Twitter´s social sensing capabilities have sufficient precision and timeliness to cater to the `extreme´ navigation needs of sports fans? And if so - how can we design a TV event detection framework that can be extended to multiple sports, and beyond sports. We focus on navigation for sports programming in the narrative, as an extremely demanding TV application that also has high market attractiveness. However, we anticipate that the ideas and architecture proposed herein apply to any TV programming that lends itself to interstitial viewing, and elicits a high level of real-time user participation in social networks.