Purpose: This study analyzes guest sentiment from TripAdvisor reviews and examines its relationship with room occupancy at the Raffles Bali Hotel. It explores key factors influencing guest satisfaction and their correlation with occupancy trends, using a time-series forecasting approach to predict future hotel performance.

Research/methodology: The research utilized TripAdvisor review data from 2020–2024, which was scraped, cleaned, and classified for sentiment using Python and Julius AI. A Seasonal Autoregressive Integrated Moving Average model was applied to sentiment data, while a SARIMAX model incorporated sentiment as an exogenous variable to forecast occupancy rates.

Results: Findings indicate that most guest reviews were positive, contributing to high overall satisfaction levels. Although occasional declines in sentiment occurred (e.g., March 2024), trends remained favorable. Time-series analysis revealed a significant positive influence of sentiment on occupancy, with slight negative short-term fluctuations. This suggests that while sentiment strongly supports long-term occupancy growth, short-term variations are less predictable.

Conclusions: Positive guest sentiment is a key driver of occupancy rates in luxury hotels. Although Granger Causality testing did not confirm a short-term causal link, long-term trends highlight the importance of managing guest sentiment to sustain occupancy levels. Hotel managers can use these insights to optimize service quality, improve guest experiences, and refine marketing strategies.

Limitations: The SARIMA models excluded external factors such as marketing campaigns, seasonal events, and competitor data. Guest demographics were not segmented in this study.

Contribution: This study introduces a novel integration of sentiment analysis and time-series forecasting, providing actionable insights to enhance service quality and improve hotel occupancy performance.