Exploratory Graph Analysis for  Well-being Composite Indicators

Traditional well-being and sustainability indicators often overlook cross-domain interactions. This study introduces a network-based framework using Exploratory Graph Analysis (EGA) to identify coherent well-being dimensions from the 2021 Italian BES dataset (NUTS-3). Utilizing EBICglasso and Walktrap algorithms, we identify eleven empirical communities that reveal latent structures not captured by predefined aggregations, including specific clusters for education and NEET indicators. The derived composite indicators show high internal consistency and confirm significant North–South geographical disparities. By integrating network psychometrics into official statistics, this paper provides a robust, data-driven approach to composite indicator construction, enhancing the interpretability of complex multidimensional systems for policy intervention.