A data-driven model for Mass Media influence in electoral context

Abstract: Mass Media outlets have occupied the central role of the political scenario, and are persuasive in the process of opinion formation of the citizens. In particular, the study of the relationship between Mass Media and behaviour of citizens can be monitored during election times, given the accessibility of news related to the candidates and polls that precede the election's day. In this paper we present a novel two-dimensional data driven Mass Media model based on semantic analysis of newspapers and national election surveys, which we use to analyse how a single influence mechanism should behave in order to reproduce the behaviour of the voters. Using simple and feasible rules for dynamics, we were able to find a notable agreement between the model's predictions and the polls which help us to understand the underlying mechanisms of the interactions between reader and media.