The current news and social media environment is increasingly characterized by the prevalence of visual content, such as memes, selfies, and news images. The large-scale availability of such data, along with the need for inductive exploration of media content, underscores the importance of integrating visuals into clustering methods. Recent studies have begun to explore opportunities to unsupervised analysis of visual content, experimenting with different data representations and clustering algorithms. However, the correspondence between clustering approaches and theoretical constructs remains unclear, enhanced by the methodological heterogeneity of existing studies. To address this issue, this study conducts a systematic literature review on approaches to clustering visual media data. Drawing on a manual analysis of N = 74 publications clustering visual media content, key concepts, clustering methods, and the validation strategies in current research are discovered. The results demonstrate gaps between theoretical and methodological advances in visual analysis and outline new strategies of bridging them together in future research. Based on these findings, guidelines for applying clustering to visual media content are developed, contributing to a more robust framework for analyzing visual content in communication science.

Ozornina, N., Zeller, A., & Haim, M. (6/2026). Making Sense of Images: A Systematic Literature Review on Approaches to Clustering Visual Media Data for Communication Science. Presented at the 76th Annual Conference of the International Communication Association, Cape Town. (content_copy)