Resumo: In recent years, different tools and frameworks have been proposed to explain the preferences of tourists when choosing accommodation in a destination so as to understand the elements that most influence this decision. This article contributes to this line of research by analysing tourist preferences in the cities of Santa Marta and Cartagena de Indias, the two most prominent destinations on the Colombian Caribbean coast. More precisely, it researches how the destinations’ attributes (tourist attractions, location, activities offered, beaches…) combined with the hotels’ features (services offered, location, price, category…) influence online comments. Both destinations are examined with the CHAID decision tree model, a tool often used in engineering but much less applied in hospitality management – although it is a highly useful one. The data collected for this research are drawn from the comments users uploaded to the platform booking.com and the results help to better characterize the two destinations and their visitors. While some of the trends observed are known to the professional community in both destinations, other latent trends are not. On a more practical note, the authors seek to prove the usefulness of big data tools in supporting decision-taking processes, particularly those depending on the knowledge of customer preferences.
