Computational Approaches to Political Dynamics: Public Opinion, Digital Platforms, and Democracy

Over the past few years, a new scientific discipline has emerged at the intersection of computer science and the social sciences: computational social science. Fueled by advances in computational methods and the growing availability of large-scale digital data, researchers from fields as diverse as computer science, physics, and mathematics have increasingly turned their attention to social and political phenomena, bringing new quantitative perspectives to their study.

In this talk, I will explore how computational tools can be used to investigate political dynamics through the analysis of behavioral data collected from digital platforms. I will first briefly introduce mathematical models of public opinion formation, before discussing how machine learning techniques, particularly dimensionality reduction, can be leveraged to systematically map online political ecosystems. I will then show how these approaches can help uncover the structure of political competition and public opinion, as well as the ways in which platform design and affordances shape these dynamics. Finally, I will broaden the discussion to consider how digital platforms reflect or distort public opinion, and what this might mean for the future health of our democracies.