This article introduces to political science a framework to analyze the content of visual material through unsupervised and semi-supervised methods. It details the implementation of a tool from the computer vision field, the Bag of Visual Words (BoVW), for the definition and extraction of “tokens” that allow researchers to build an Image-Visual Word Matrix which emulates the Document-Term matrix in text analysis. This reduction technique is the basis for several tools familiar to social scientists, such as topic models, that permit exploratory, and semi-supervised analysis of images. The framework has gains in transparency, interpretability, and inclusion of domain knowledge with respect to other deep learning techniques. I illustrate the scope of the BoVW by conducting a novel visual structural topic model which focuses substantively on the identification of visual frames from the pictures of the migrant caravan from Central America.

This study focuses on the Hong Kong Lennon Walls and the communications posted there. We assert that the physical placement of COVID-19 related images on the Lennon Walls of Hong Kong and the replication of symbols and iconography from the Umbrella Movement and the Anti-ELAB Movement situated COVID-19 discourse not only physically within but also symbolically within the contentious politics of Hong Kong. We conclude that the messages and images posted on Lennon Walls between January and April 2020 have used COVID-19 to extend public expression of sentiment on the debates around the Hong Kong government and to further mobilize a sense of Hong Kong identity against China. The findings contribute to the understandings of how the cultural politics surrounding the pandemic became a collective action frame in the mobilization of a localized Hong Kong political identity against the Hong Kong and Chinese governments.

Hydraulic fracturing (HF) is a focal topic in discussions about domestic energy production, yet the American public is largely unfamiliar and undecided about the practice. This study sheds light on how individuals may come to understand hydraulic fracturing as this unconventional production technology becomes more prominent in the United States. For the study, a thorough search of HF photographs was performed, and a systematic evaluation of 40 images using an online experimental design involving $N=250$ participants was conducted. Key indicators of hydraulic fracturing support and beliefs were identified. Participants showed diversity in their support for the practice, with 47 percent expressing low support, 22 percent high support, and 31 percent undecided. Support for HF was positively associated with beliefs that hydraulic fracturing is primarily an economic issue and negatively associated with beliefs that it is an environmental issue. Level of support was also investigated as a perceptual filter that facilitates biased issue perceptions and affective evaluations of economic benefit and environmental cost frames presented in visual content of hydraulic fracturing. Results suggested an interactive relationship between visual framing and level of support, pointing to a substantial barrier to common understanding about the issue that strategic communicators should consider.