Analysing and understanding tourism mobility constitutes a tremendous challenge for states, cities and companies involved in cultural heritage. Traditional methods of tourist flow analysis often rely on survey data and observational studies to map out general trends and movements, focusing on aggregated visitor numbers and basic demographic information without delving deeply into the complex interconnections and sequential patterns of tourist behaviours. This study introduces SeqPatTour, a comprehensive methodology that innovatively mines and analyses tourist behaviour through sequential pattern analysis with data from Tripadvisor. Distinguished by its ability to perform both quantitative and qualitative analysis, SeqPatTour navigates the complexities inherent in analysing sequential patterns on graphs, facilitating a deeper understanding of tourist movements. It critically assesses path extraction and sequential pattern mining techniques to tailor the analysis to specific temporal constraints, thereby introducing specialised metrics for the tourism domain. Our methodology, applied on the city of Lille with 287,429 reviews, leading to a graph of 3,317 nodes and 94,829 relationships, has the potential to enhance analytical capabilities in tourism studies, providing actionable insights for urban planning and policy formulation. Its approach to integrating path extraction with temporal-gap constrained sequential pattern mining effectively uncovers latent behavioural patterns among tourists. This not only underscores the method’s value in improving tourist experiences but also its contribution to economic development in tourism-focused areas, navigating through the intricacies of sequential pattern analysis on graph structures to offer profound insights into tourist dynamics.
