Abstract:
The overarching goal of this paper is to gain visibility on tourist preferences and whether or not the needs of tourists are being met. With the Travel and Tourism (T&T) sector being the backbone to the global economy and the sector becoming more saturated and competitive, insights on T&T are vital now, more than ever. The rise of social media and user-generated content has effectuated the opportunity for a systematic analysis of tourist preferences via user-generated content. This paper is focused on gaining insights into tourism in Bangkok, Thailand through user-generated content scraped from TripAdvisor’s online reviews of tours and activities. In order to develop insights on tourist preferences and tourism trends in Bangkok, various analyses were implemented, including sentiment analysis to gather tourist point-of-view, association rules mining to find patterns of preferences, and natural language processing along with text frequency analysis to understand what features tourists are most frequently talking about. This paper also developed machine learning prediction models using Logistic Regression, Support Vector Machine, and Random Forest algorithms to forecast 5-start ratings and 1-star ratings of reviews – with the purpose of identifying factors that significantly affect positive and negative sentiments on tours/activities.
