Knowing a Good Show When You See One

Social media has become an integral part of the web, and its popularity continues to provide an outlet for people´s opinions and discussion about any topic of interest. In this paper we examine the interest around a number of television series that are broadcast on a weekly basis. Our aim is to show that by observing solely the initial interactions of fans or users of a web forum, we can extrapolate the longer-term interest in particular episodes. We do so by observing the temporal dynamics of the conversation, and performing a clustering so as to judge how much time is required before reasonable conclusions can be drawn about the level of interest surrounding an episode. We find that early interaction trends have strong similarities with the overall conversation patterns.