Cultural tourism has evolved into a mass market phenomenon that contributes a sizeable portion to international tourist arrivals in Europe. Yet, exact estimates of cultural tourists are hard to come by, due both to a lack of standardized conceptualization, and a difficulty in operationalization. Mostly, estimates are based on visitor surveys, which are expensive to conduct, infrequent, and often do not allow an in-depth analysis of the phenomenon. This paper proposes an alternative analytical methodology, scraping user-generated content and applying computational text analysis and association rule mining on visitor reviews in order to establish both centrality of cultural travel motives and improve understanding of cultural tourism typologies via analysing topical associations within the reviews. The methodology is tested on 2507 reviews for the historical centre of the city of Ghent, Belgium. The results show estimates that are comparable in size to visitor survey statistics, while lending additional information on relative importance of cultural travel motivations.