Advanced multidimensional spatial clustering for sustainable transport assessment

Nowadays, there is a growing interest in measuring the level of sustainability of transport of geographical areas. The aim of this paper is to provide a spatial clustering, defining a sustainable transport planning on the basis of environmental, social, and economic aspects. A high number of indicators is considered as a new set of cluster factors, thus groups of homogeneous provinces, taking into account the heterogeneity of the considered factors, are identified. In this paper, a novel procedure that combines hierarchical cluster and geographically weighted principal components analysis (GWPCA) is proposed to evaluate the underlying factors of the variables, without neglecting the spatial dependence. In particular, it is shown how an unsupervised method can be used to generate hierarchical clusters of variables by using correlation as distance metric. Then, the variables within each cluster are included as the input covariates of GWPCA to derive a composite spatial indicator. The results obtained by using this novel procedure are illustrated through a real dataset characterized by socio-economic indicators of transportation sustainability for the Italian provinces with reference to the year 2022.