Protest Behavior

Research on protest behavior increasingly leverages advanced computational methods to analyze large datasets of textual and visual information from various online platforms. Current studies employ deep learning techniques, including transformer models like Longformer and Swin-Transformer, to automatically identify and classify protest events within news articles and social media images and videos, often focusing on the visual framing of protests. This work aims to improve understanding of protest dynamics, particularly how different media platforms shape public perception and participation, and offers valuable tools for social scientists studying collective action.