Functional Graphical Models to map Brexit debate on Twitter

In recent years a literature on multivariate functional graph models has been developed. The graphical representation of the conditional dependence among a finite number of random variables is indeed appealing in different applications, such as e.g. the analysis of the brain connectivity. We want to investigate a novel extension of this methodology, considering random functions spatially and temporally correlated. A motivating case study is the analysis of the semantic network that tracks the change of the Brexit debate on Twitter across UK during a particular time frame. By considering the change in time of a word usage as a functional realization, a semantic network on the topic of interest is defined by a graphical representation of the conditional dependence among functional variables.