This project investigates how conspiracy theories and extremist rhetoric are subtly introduced and normalized through popular media. Using transcripts from over 180 episodes of the History Channel series Ancient Aliens, structural topic modeling (STM) was applied to trace how pseudoscientific and mythological storytelling evolves over time into themes often associated with alt-right and xenophobic discourse.
The project was originally developed during my PhD work at UNC Chapel Hill to explore how seemingly innocuous entertainment content can serve as a vector for ideological influence, particularly through repetition, legitimization of pseudohistory, and appeals to anti-elite sentiment.
- Collected and cleaned 180+ full episode transcripts
- Implemented STM models using both VEM and Gibbs sampling to uncover latent thematic structures
- Conducted longitudinal topic tracking to analyze how rhetorical themes shift seasonally
- Developed an interactive RShiny app for public exploration of model results, topic proportions, and representative episode texts
Tools & Stack
R · RShiny · STM (Structural Topic Modeling) · Tidytext · NLP · Text Mining · VEM · Gibbs Sampling
Live Demo: jonathanschlosser.shinyapps.io/AncientAliensApp
GitHub Repository: JPSchloss/Ancient-Aliens