Investigating the Relationships among Sentiment, Hotel Aspects, and Customer’s Home Country from Online Reviews: A Machine Learning Approach

Abstract

Comments of travellers about different characteristics of a hotel are one of the most important inputs for managers to enhance their services, facilities and their marketing campaigns. In finding a way to improve the practical experience of both buy-side and sell-side in the hospitality market, we apply sentiment analysis for hospitality data from an user-generated content site which is TripAdvisor. Typically, from both quantitative data and qualitative data including reviewer's comments, our contributions are first to propose a framework to uncover aspects/features of a hotel that customers are focusing and second to measure traveler's sentiment toward different aspects. By a large-scale analysis of 410 hotels in Ho Chi Minh city, we add more contributions to finding the emerging opinions of customers towards different aspects of the hotel's reviews at both the macro and micro level. Specifically, the performance in terms of sentiment towards key aspects is stable over time and the gaps between different hotel groups are clearly captured by our framework. The aspect-to-aspect comparisons for individual hotels show the strengths and weaknesses of their services. Finally, the analysis reveals that customers from developed countries and developing countries show distinct patterns. These findings play important roles in assisting hotel management practice and in understanding customer perception towards hotel services.