COSMOS Research Center is pleased to announce a recent publication in Springer’s Journal of Social Network Analysis and Mining titled “Discovering Cross-Platform Narrative Flow Templates Using Frequent Subgraph Mining,” authored by Ridwan Amure and Nitin Agarwal.

Social media narratives rarely remain confined to a single platform. They often move across digital ecosystems, adapting to the unique affordances of each space, from long-form commentary on YouTube to short-form expression on TikTok, public discourse on X, and visual storytelling on Instagram. This study addresses the need to better understand how narratives evolve across platforms over time.

The paper introduces a graph-mining framework for discovering narrative flow templates, which are recurring structural patterns that describe how narratives propagate between platforms. By adapting the SoPaGraMi algorithm for frequent subgraph mining, the study moves beyond traditional sequence-based approaches and captures more complex, non-linear diffusion structures in cross-platform discourse.

Using the recent Tariff War as a case study, the research analyzes narrative activity across YouTube, TikTok, X, and Instagram. The framework uncovers distinct diffusion motifs, including platform pathways associated with pro-American, pro-Chinese, and cooperative US-China narratives. These findings show how different platforms can serve complementary analytical, expressive, and visual functions in shaping online discourse.

This work contributes to the growing field of computational social science by offering a scalable approach for identifying coherent cross-platform diffusion structures. It advances COSMOS’s broader efforts to understand how narratives spread through multimodal social ecosystems and how data-driven methods can support deeper analysis of online information dynamics.

Read the full article here.