CLUSTERING OF TOURISM PATTERNS WITH SELF-ORGANIZING MAPS: THE CASE OF SICILY

Abstract

In this paper the Self-Organizing Map algorithm is used for studying whether and which tourism flows in Sicily are synchronized, i.e. which flows show similar patterns in time and space, if any. Synchrony hunting was performed for domestic and international tourists both on a yearly and monthly basis. Local tourism, meaning the holidays spent in Sicily by residents in the island, is also considered but on a yearly basis only. The analysis makes use of time series representing the number of overnight stays in Sicily over the period 2013-2019. Results provide evidence for a domestic market overall more synchronized than the international one, both in time and space. Spatiotemporal patterns for local tourism seem strongly influenced by the landscape. Some policy implications are drawn.