Automated Multimodal Analysis

About

This course introduces computational methods for analysing visual and multimodal data in social science research. It is designed for PhD students and early‑career researchers across disciplines – including political science, sociology, communication, and public policy – who work with image‑rich digital data. Basic familiarity with Python or R is assumed.

Key topics include automated image classification, unsupervised visual analysis, object and face detection, and multimodal approaches that combine images with text, including CLIP‑style models and larger vision–language models (VLMs). Methods are grounded throughout in real datasets and linked to substantive research questions across the social sciences.

The course combines pre‑recorded introductory videos and readings with hands‑on coding exercises across online and in‑person sessions. A project thread runs through the full course, ending in a short presentation. You may also submit a post‑course written analysis for an additional ECTS credit.

Course leader

Felicia Loecherbach is Assistant Professor of Political Communication and Journalism at the University of Amsterdam and Faculty Research Affiliate at NYU's Center for Social Media and Politics. Her research focuses on news diversity, visual social media, and algorithmic curation, with expertise in computational methods and automated content analysis. A core strand of her work involves the automated analysis of political images at scale, examining how visual content shapes political communication on social media platforms. She holds a PhD in Communication Science and has extensive experience teaching graduate-level courses on big data, computational methods, and data journalism.

Target group

Social science researchers

Course aim

This course aims to equip participants with the skills to apply computational methods for visual and multimodal analysis in social science research, using techniques from computer vision and natural language processing.

Fee info