Analyzing and Predicting Housing Affordability Trends in the UK Using Machine Learning

Varghese, Ebin; Muyeba, Maybin K; Mohammadi, Azadeh

Abstract

Housing affordability has become an increasingly critical issue in the United Kingdom, driven by rising house prices and rent levels that outpace income growth. This study leverages machine learning and data-driven techniques to analyse and predict affordability trends across both ownership and rental markets. We utilize time series forecasting (Prophet) to project future house prices at national and regional levels, apply K-Means clustering to segment regions by rent-based affordability, and develop a Random Forest classifier to predict where individuals can afford to rent based on salary. Drawing from diverse public da-tasets, including housing prices, rent data, income, inflation, and mortgage rates, we construct engineered features such as affordability indices and rent-to-income ratios. Our findings reveal significant spatial disparities in affordability, strong urban-rural contrasts, and persistent affordability stress in high-demand areas such as London and Manchester. The models provide predictive insights for housing planners, policymakers, and tenants navigating a volatile housing landscape

• Presentation Conference Type: Conference Paper (published)
• Conference Name: International Conference on Data Science, AI and Applications
• Start Date: Jul 19, 2025
• End Date: Jul 20, 2025
• Acceptance Date: Jun 22, 2025
• Deposit Date: Aug 5, 2025
• Print ISSN: 0302-9743
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Keywords: Housing Affordability; Machine Learning; Forecasting; Clustering; Classification; UK Housing Market
• Publisher URL: https://www.springer.com/gp/computer-science/lncs