Computational Thematic Analysis of Academic Representations of the Gisaeng in the Korea Citation Index (2000–2024)

This repository contains the dataset and code used for the article “Computational Thematic Analysis of Academic Representations of the Gisaeng in the Korea Citation Index (2000–2024)”. The study applies natural language processing and topic modelling to 613 abstracts published between 2000 and 2024 in South Korean academic journals indexed in the Korea Citation Index (KCI). Using Ko-SRoBERTa embeddings, UMAP dimensionality reduction, and HDBSCAN clustering implemented through BERTopic, the analysis identified ten thematic clusters, subsequently organized into four macro-themes: Literature, Performing Arts, Colonialism and Modernity, and Iconic Gisaeng. The repository includes: • The curated dataset of 613 abstracts (.csv). • Jupyter Notebook scripts for preprocessing, modelling, and visualization. • The Dash/Plotly application files (app.py, requirements.txt, Dockerfile, and processed dataset) used to generate the interactive UMAP visualization. • A link to the interactive version of the UMAP visualization hosted on Hugging Face Spaces: https://huggingface.co/spaces/Cesarhanmun/UMAPGisaengKCI. This deposit is intended to ensure transparency, reproducibility, and long-term accessibility of the data and code. It can be cited via the DOI provided by Zenodo.