Territorial Attractiveness for Foreign Population through a Spatio-Temporal Blind Source Separation Model

Nowadays, understanding the settlement patterns of foreign residents and the consequent territorial attractiveness of municipalities has become increasingly important for the overall life quality assessment. Although numerous studies analyzed and predicted the presence of foreign population across territories, they did not jointly consider the spatial and temporal dimensions of the potential impacting factors. This paper proposes an innovative predictive approach which combines the spatio-temporal blind source separation methodology to identify latent components of territorial attractiveness to foreigners, with geostatistical estimation techniques to model the dependence structure in space and time of the phenomenon under study. The procedure is applied to a multivariate dataset, covering 7,606 Italian municipalities over the period 2001–2021 and including demographic and economic indicators, based on official statistics, while data from 2022 to 2025 are reserved as a test set for out-of-sample forecasting. A comparison with a traditional time series model demonstrates that the proposed framework significantly improves predictive performance, confirming the importance of jointly modeling spatial and temporal dependencies. Finally, the analysis of variance and the spatial Gini index are used to assess inequalities across municipalities. The findings provide useful insights for understanding territorial attractiveness and for supporting place-based demographic and migration policies aimed at improving the quality of life of residents across Italian municipalities.