Hotel online reviews: What influences the sentiment and rating? A sentiment analysis on Lisbon hotels

Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Information Systems and Technologies ManagementThis study analyzes 38,292 reviews from 192 Lisbon Hotels posted in Booking.com and TripAdvisor for hotels located in Lisbon, Portugal. The aim of this research is to understand if the hotel categories and nationality of the reviewer influence the sentiment and the ratings of the reviews. Moreover, it also evaluates whether the review ratings in Booking.com are inflated in comparison with TripAdvisor. A sentiment analysis on all the reviews was performed using Semantria, a software from Lexalytics. The key insights of this research are that the hotel category influences the sentiment on the review, even though the sentiment does not increase as the hotel category increases. Furthermore, this study also found statistically significant differences between the sentiment depending on the nationality of the reviewer. This work also rejects the hypothesis that the review ratings on Booking.com are inflated in comparison with TripAdvisor. Finally, this study provides insights to hotel managers who should focus on meeting the customers’ expectations and adapt the service provided according to the nationality of the reviewer. Moreover, by not providing evidence that Booking.com review ratings are inflated, this study restores the confidence of the stakeholders in the reliability of the scoring system applied by that platform