The structure of the network of visited attractions in an urban destination

Abstract

Large destinations such as cities usually offer multiple and diverse attractions and the tourists visit a selection of them during the stay. The set of co-occurrences of visits can be represented through a weighted network where the nodes are the attractions and the edges represent the number of tourists that visits two different attractions. This study analyzes the factors influencing the co-occurrence of visits to attractions in the city of Madrid. Starting from data of visits extracted from user comments in TripAdvisor along five years (2013-2018), we build the network of attractions in the city and apply exponential random graph models to estimate which factors influence the link between two different attractions. The temporal trend of this network of co-occurrences is also analyzed. The results show that the tourist decision to visit two attractions in the same trip is influenced by their popularity and other relational characteristics, such as the type of attraction. The identification of the forces that connect attractions in this network is crucial for an efficient management, promotion and preservation of them.