Analyzing Political Polarization Through Natural Language Processing and Online Communities
Political polarization poses a significant threat to modern democratic societies. It weakens public discourse, erodes social trust, and exacerbates hostility among citizens, ultimately endangering the democratic decision-making process. While numerous attempts have been made to analyze political polarization, recent developments have opened new avenues of research. The rise of platforms such as online communities has enabled the collection of political opinion data from public, while advancements in natural language processing have facilitated more sophisticated analysis of textual data.
This study analyzes the state of political polarization in South Korea by utilizing data collected over a specific period from Korean online political communities. The analysis employs methodologies including tokenization, word2vec embeddings and sentiment analysis.