#2024election: An Exploratory Analysis of Politicultural Communities on TikTok

This thesis explores how cultural discourse varies in both structure and content across political party lines on TikTok. To answer this question, I take presidential candidate accounts as partisan representations and examine the TikTok videos reposted by followers of candidate accounts. To analyze the TikTok repost data, I employ frequent itemset mining, social network analysis, and BERTopic modeling. I find that the semantic communities in Trump followers' reposts are more overlapping and less distinct than those found in Harris followers' reposts. I find that Harris followers' reposts cover a wider breadth of topics. I analyze several emergent themes in the content of the reposts, finding connections to gender, sexuality, and violence. These findings corroborate current literature surrounding American partisanship, culture, and social media. I discuss how TikTok mediates and produces culture, indicating that future research is needed to understand the political relevance and impact of the social media platform.