Automated Image Analysis in Practice

Session Submission Type: Full Paper Panel

Session Description

Automated image analysis holds great promise for political science research. This panel, one of two linked panels using deep learning for the analysis of big data, focuses on applications of machine learning methods for image data. The authors address a range of substantive questions that demonstrate the potential for computer vision to expand our understanding of political phenomena. Pan and Lu use deep learning to classify mobile phone screenshots from China to examine the effects of censorship. Casas, Webb Williams, Aslett, and Wilkerson ask whether images affect online support for social movements, comparing the image content of successful and unsuccessful mobilizations with an unsupervised image clustering technique. Homola and Torres analyze newspaper coverage of the Women's March with a visual topic model to better understand how the media frames contentious events. Similarly, Boxell is also interested in understanding media biases, analyzing roughly 8 million images to examine how politicians are portrayed. In sum, the panel demonstrates how images as data can unlock questions on censorship, protest, and media framing.

Sub Unit

• Division 8: Political Methodology

Individual Presentations

• Deterrence and Censorship - Jennifer Pan, Stanford University; Yingdan Lu, Stanford University

• The Shades and Shapes of the Pink Wave: Visual Perspectives of the Women's March - Jonathan Homola, Rice University; Michelle Torres, Rice University

• Slanted Images: Measuring Nonverbal Media Bias - Levi Boxell, Stanford University

• Is the Picture in Focus? Images in Social Movement Mobilization - Andreu Casas, Vrije Universiteit Amsterdam; Nora Webb Williams, University of Illinois at Urbana-Champaign; Kevin Aslett, New York University; John D. Wilkerson, University of Washington

Chair

• Jungseock Joo UCLA

Discussants

• Han Zhang The Hong Kong University of Science and Technology

• Bryce Dietrich University of Iowa