University of Twente Student Theses
As of Friday, 8 August 2025, the current Student Theses repository is no longer available for thesis uploads. A new Student Theses repository will be available starting Friday, 15 August 2025.
Election-Driven Trends in Reddit Hate Speech : A Clustering Approach to Target Analysis Before and During Trump Campaign Periods in USA
Andrei, Rebecca (2025) Election-Driven Trends in Reddit Hate Speech : A Clustering Approach to Target Analysis Before and During Trump Campaign Periods in USA.
|
PDF
657kB |
| Abstract: | The intersection of political campaigns and online discourse is critical for understanding the emergence of hate speech. This paper examines how Donald Trump’s 2024–2025 presidential campaign influenced hate speech patterns on Reddit, a major platform for political discussion. Using 55M Reddit posts collected before and during Trump’s campaign, we apply Natural Language Processing (NLP) techniques to analyze the dynamics of hate speech over time. Hate speech classification is performed with a fine-tuned Bidirectional Encoder Representations from Transformers (BERT) model, which captures deep contextual meaning and detects subtle, implicit forms of hate speech common in political conversations. Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) is then employed to identify latent groupings of hate speech targets. HDBSCAN is chosen over traditional DBSCAN for its ability to handle varying cluster densities, noisy data, and high-dimensional embeddings produced by BERT. Cluster boundaries are further refined using an algorithm that separates overlapping clusters to delineate target categories clearly. This approach offers empirical insights into how political campaigns may shape toxic online behavior and broader trends in social polarization, informing academic research and policy discussions on digital discourse, political communication, and platform moderation. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 50 technical science in general, 54 computer science, 77 psychology, 88 social and public administration, 89 political science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/107506 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page