An Innovative Way to Model Twitter Topic-Driven Interactions Using Multiplex Networks

We propose a way to model topic-based implicit interactions among Twitter users. Our model relies on grouping Twitter hashtags, in a given context, into themes/topics and then using the multiplex network model to construct a thematic multiplex where each layer corresponds to a topic/theme, and users within a layer are connected if and only if they used the same hashtag. We show, by testing our model on a real-world Twitter dataset, that applying multiplex community detection on the thematic multiplex can reveal new types of communities that were not observed before using the traditional ways of modeling Twitter interactions.