02.3 - Development of a Predictive Tool for Real Estate Analysis Using Machine Learning Techniques

• Ricardo Reier

  Catholic University of Ávila

  Author

• Gregorio Acedo

  Catholic University of Ávila

  Author

• Luis Miguel Garay

  International University of La Rioja

  Author

• Sergio Náñez

  Catholic University of Ávila

  Author

Housing markets in many countries are currently facing severe affordability challenges, particularly in large urban areas where prices have risen faster than wages. In Spain, and especially in cities such as Madrid, housing prices have reached levels comparable to the 2007 bubble, intensifying concerns about access and urban inequality. This study develops a predictive tool for real estate valuation based on Big Data and Machine Learning techniques. Using automated data collection, spatial analysis, and Gradient Boosting algorithms, the proposed system estimates property market values in real time by integrating structural and georeferenced variables. Madrid is selected as a case study due to its size, economic relevance, and highly dynamic housing market. The results demonstrate that machine learning models effectively capture intra-urban price heterogeneity and outperform traditional valuation approaches in predictive accuracy. The study also presents an interactive application that translates academic research into a practical decision-support tool for buyers, sellers, investors, and policymakers. By combining methodological rigor with applied relevance, this research contributes to the literature on PropTech and real estate analytics, while highlighting both the potential and the limitations of AI-based tools in addressing structural housing market challenges.