Predicting social welfare in Madrid neighbourhoods using machine learning

This research aimed to identify the principal factors predicting social welfare and inequality in the neighbourhoods of Madrid city. A comprehensive dataset representing various socioeconomic metrics of Madrid’s neighbourhoods is analysed utilising different linear regression models and the XGBoost machine learning algorithm. The findings indicate that demographic variables play a crucial role in shaping social welfare and inequality in Madrid's neighbourhoods, with the percentage of women, and the percentage of children under 14 years old and adults over 65 years old being the most important variables for predicting social welfare and inequality in the studied neighbourhoods. Contrary to expectations, the average net income per household does not emerge as a significant predictor. Policymakers and urban planners should consider these factors to ensure equitable development and welfare distribution. The low impact of average net income per household on social welfare and inequality observed in this study should be regarded with some caution, but our findings certainly underline the need for a deeper understanding of socioeconomic dynamics in urban settings.