Visual content dominates the current digital environment, yet there are still some methodological issues concerning the analysis of symbolic elements in the videos. This paper fills this gap by exploring frame extraction techniques and quantifying symbolic transmission dynamics across social media platforms. After comparing the different extraction methods, we observed that the standout frames, which are the points of significant visual change, contain significantly more symbolic content than the first frames or the thumbnails, which provided 59% more symbols in TikTok and 39% more in YouTube. To determine how these symbols are spread, five epidemiological models were used to measure the transmission rates of the different symbolic elements: SIS, SIR, SIRS, SEIR, and SEIZ. The results showed completely different platform-specific patterns; for instance, cultural symbols had the highest transmission rate in TikTok under the SIRS modeling. On the other hand, YouTube was most compatible with political symbols and content with any two symbolic elements only. These findings shed light on how the effectiveness of a narrative depends not only on the symbolic content provided but also on the architecture of the platform and the optimal number of symbols. By combining an approach to the extraction of visual data with epidemiological modeling, we offer both useful strategies for analyzing video content and a deeper understanding of information diffusion on contemporary digital networks.

PDF