Decoding the Popularity of TV Series: A Network Analysis Perspective

Viewer ratings are a key factor in shaping the success and continued viability of television shows. However, the factors contributing to high or low ratings are complex. This paper investigates whether character interactions in TV episodes, as captured through character network analysis, correlate with viewer ratings. Character networks are graphs created from the plot of a TV show that represent the interactions of characters in scenes. We constructed character networks from episode plots of three popular TV series and extracted graph metrics such as network density and centrality measures from these networks. We explored the hypothesis that interactions between the characters in a TV episode, as quantified by the metrics of the episode character networks, are related to the overall reception of the episode, as indicated by the TV episode reviews. We conducted regression analysis between character network metrices and the TV reviews. The findings revealed significant, albeit varied, correlations between certain network metrics and episode reviews across the different series. For instance, Game of Thrones episodes with fewer active characters received higher reviews, suggesting viewer preference for simpler character interactions. Similarly, Breaking Bad also gained popularity by centering on a smaller group of main characters. Our study contributes to understanding how character dynamics influence TV reviews, providing valuable insights for TV producers and writers in developing their shows. We have shown that network analysis has the capability to quantify character interactions. This could potentially allow producers to consider adjusting character dynamics in future episodes, which may enhance audience engagement.