Seeing the Unseen? Opportunities and Challenges in Understanding Tourist Behaviors from UGC using Large Language Models (LLMs)

Gusriza, Fondina
Priyambodo, Tri Kuntoro
Mustofa, Khabib
Mutiarin, Dyah

International Journal of Informatics, Information System and Computer Engineering (INJIISCOM)
Vol. 8 No. 1 (2027): INJIISCOM: VOLUME 8, ISSUE 1, JUNE 2027 (Online First)

The era of digital tourism has resulted in various of User-Generated Content (UGC), such as online reviews, social media, and travel blogs. UGC is an important source of data for understanding tourist behaviours but often overlooked. This study explores the opportunities and challenges of using Large Language Models (LLMs) as a new approach to analyse UGC. The data were collected from TripAdvisor, focusing on 15 cultural heritage sites in Yogyakarta, Indonesia, with more than 13,000 user reviews. OpenAI’s Large Language Model was applied through API integration in Python, and a Natural Language Processing (NLP) approach was used to analyse tourist motivations, experiences, and satisfaction. The analytical process includes web scraping, text pre-processing, theory-driven classification, prompt engineering, and lexicon mapping. This paper contributes to the emerging discourse on the use of Artificial Intelligence in tourism studies, particularly in understanding tourist behaviours through textual data.