Leveraging Data Analytics to Manage Visitor Flows in Heritage Sites

This paper presents a data-driven methodological workflow for the analysis of visitor flows, applied to the architectural and landscape system of Pitti Palace and Boboli Gardens in Florence (Italy). This heritage site combines the periodic monitoring of stone façades – prone to deterioration and detachment – with high visitor pressure, as millions of people visit its five museums each year. The method relies exclusively on raw data already available to site managers, namely tabular records from the ticketing system, thereby ensuring transferability and privacy protection without visual user tracking. The analysis covers 18 months in 2024- 2025 and begins with the selection and filtering of relevant variables, including entry times to the museums and ticket categories. The resulting dataset is processed through big data approaches combining unsupervised Machine Learning and statistical analysis, with a focus on identifying homogeneous clusters by overall attendance, seasonality, and user categories. For each cluster, temporal distribution analyses enable the identification of trends and outliers. The workflow can then be extended to infer visit sequences and dwell times by integrating on-site measurements of access durations and travel times among the site’s museums, enabling the reconstruction of flow dynamics across the complex. The proposed workflow is intended to support site managers in decision-making processes to optimise visitor flows and address heritage conservation issues, while enhancing user safety and comfort.