Guest talk: Matteo Magnani "From Pixels to Meaning: Making Sense of Online Political Communication at Scale"

From Pixels to Meaning: Making Sense of Online Political Communication at Scale

Matteo Magnani
Uppsala University

Abstract: From viral memes to visual depictions of protests and societal issues, images play a central role in online political discourse. Images are powerful in triggering emotion, which can drive human behaviour, and they can cross linguistic and (to some extent) cultural boundaries better than text. Yet, while computational analysis of text is well established, the visual dimension has only recently begun to attract systematic attention. In this talk, I will explore how computer vision is being used and adapted for the study of political communication: what makes social image analysis distinct from other applications, and how its specific objectives influence the development of methods to identify clusters, topics, and narratives in visual data. I will also reflect on the potential misuses of these technologies, and what it will take to ensure they advance rather than compromise communication research.

Bio: Matteo Magnani is Professor of Data Science and Head of the Computing Science division at Uppsala University, where he leads the Uppsala Information Laboratory and teaches Social Data Mining. His research is at the intersection between computer and social science, with a focus on social network analysis (on social media and population registers), visual political communication, and social cybersecurity.

This guest talk is hosted by Associate Professor Mikko Kivelä, Department of Computer Science.