Applying Automatic Text Analysis Methodologies to Audiovisual Serial Product

Thegoalofthispaperistoexploretheapplicationofautomatictextanalysismethodologiestocontemporaryaudiovisual serial narratives. As a case study we use the Apple TV+ seriesServant(Apple TV+, 2019-). Wefirst focused on theprimarytext(the English dialogue used in the TV series) to examine the role of thedialogue and the characters’ interactions through an exploratory application of Social Network Analysis.In particular, tagging the dialogue in XML format allowed the identification and quantification of scenes,characters and speaker-receiver pairs that were used to implement Social Network Analysis. Secondly, wecollected tweets as thesecondarytext(the text produced by the Twitter audience), and analysed users’behaviours and preferences considering both semantic and quantitative points of view. We underlined howanalyses conducted on tweet sentiment can help to monitor this social engagement mechanism and howit may evolve over time. The paper is highly experimental in that, in addition to findings related to thenarrative structure of the serial product (thanks to theprimarytext) and analysis of the relationship overtime with the audience (thanks to thesecondarytext), it aims to test a shared analytical framework that canenable large-scale comparative investigations of contemporary TV series.