O fabulosíssimo projeto de pesquisa de Jermain Kaminski e Michael Schober é resumido no artigo Social Networks in Movies. Segue o abstract:

Although humanities and cultural studies have a long tradition in formalistic interpretations of works of art and literature (e. g. Wellek and Warren, 1956), only few writers have understood these works as networks of characters that (inter)act with each other. This paper expands this stream of thought by extracting, visualizing and analyzing the evolution of the characters’ interaction networks in about 797 movies from 1915 to 2011 and teasers a collobarative online solution of the technique, which is crowdsourcing the further development. There are only a very few examples of research where the social network of movies has ever been analyzed. These studies mostly had a low scale (n=1) and were not automated with machine learning algorithms (e.g. Hillman, 2011; Park et al., 2011; Ding and Yilmaz, 2010). Looking at the amount of analyzed movies and technique, this paper is a new approach.

Além de apresentar esta interessante perspectiva na interpretação de obras artísticas, os autores lançaram o site moviegalaxies, que apresenta as redes sociais de personagens de filmes. As quase 800 redes podem ser visualizadas de forma interativa, com os agrupamentos, métricas de grau, intermediaridade e cluster. Abaixo alguns exemplos de redes: Magnolia (diversos grupos, resultado das diversas linhas narrativas), Solaris (rede pequena e densa do sci fi claustrofóbico, que se passa em uma nave espacial), e Fight Club (rede altamente centralizada em Jack, protagonista narrador em primeira pessoa).

Magnolia

Solaris

Fight Club