Unlocking the Power of Data Analytics in the Real Estate Market: Evidence from Italy

This thesis investigates the Italian residential real estate market through a data-driven approach, examining how macroeconomic conditions, demographic dynamics, institutional factors, and property-level data influence house prices, rental trends, and investment opportunities. Particular attention is given to regional disparities and to the different housing market trajectories of Milan, Venice, Bologna, and Rome. The study combines insights from housing economics, econometrics, and machine learning, drawing on hedonic pricing theory and recent research on automated valuation models (AVMs). Using a quantitative framework implemented in Python and R, the analysis integrates official data from OMI (Agenzia delle Entrate), ISTAT, Banca d’Italia, and Eurostat, as well as web-scraped listing data for selected Italian cities. The methodology includes descriptive statistics, data visualisation, ARIMA time-series forecasting, hedonic regression models, and Random Forest algorithms for valuation, mispricing detection, and investment screening. The results highlight strong structural heterogeneity in the Italian housing market, including persistent regional differences, contrasts between metropolitan and smaller markets, and price premiums for new-build properties. At the city level, Milan shows the strongest and most sustained price growth, Bologna demonstrates stable and resilient performance, Venice reflects marked micro-locational effects, and Rome exhibits high spatial price dispersion. From a methodological perspective, Random Forest models generally outperform linear hedonic regressions in predictive accuracy, while hedonic models remain valuable for interpretation and economic explanation. To sum up, the thesis shows how data analytics can support more informed housing and investment decisions for households, investors, and policymakers, while also acknowledging data limitations and the importance of contextual market interpretation.