Search

A key issue in tourism management relates to the lack of consensus regarding a theoretical and practical definition of the term "tourist." In turn, this results in a range of methods for counting tourists and measuring tourism. This paper presents a novel non-linear model for classifying international tourists in urban settings, based on machine learning classification methods. These methods utilize innovative feature engineering derived from photos posted on the Flickr social media platform combined with the specific urban destination street structure. The data science model that we developed for identifying international tourists produced an overall accuracy of 69% for Manhattan and 94% for Vienna and Prague, offering new tourism indicators such as repeat visits, travel distances, and short stays. The outcome of this study offers a better understanding of travel patterns among international tourists, which could improve international tourism management and promote a more practical and adaptable model for measuring and analyzing international tourism using machine learning and user-generated content. [ABSTRACT FROM AUTHOR]