The digital tourism era generates vast User-Generated Content (UGC), offering critical yet underutilized insights into traveler behavior. This study explores the opportunities and challenges of utilizing Large Language Models (LLMs) to analyze unstructured UGC. Drawing data from TripAdvisor, we examined over 13,000 reviews across 15 cultural heritage sites in Yogyakarta, Indonesia. Through a Python-integrated OpenAI API, a Natural Language Processing (NLP) framework was deployed to evaluate tourist motivations, experiences, and satisfaction. The methodology encompasses web scraping, text pre-processing, theory-driven classification, prompt engineering, and lexicon mapping. This paper bridges classical tourism theories with advanced AI, contributing to the emerging discourse on computational behavior analysis in digital tourism research.
Copyrights © 2027