AI-Driven Investment Property Recommendations Using Spatial Big Data, Price Trends, and Amenity Mapping

Abstract

In the real estate domain, investment decisions rely heavily on spatial and economic context, yet most digital platforms still provide static listings with limited personalization or geographic intelligence. The primary objective of this paper is to introduce and validate a spatially enriched recommendation system for real estate investment that integrates Artificial Intelligence (AI), Geographic Information Systems GIS), and big data analytics. Evaluated on over 70,000 property listings, the system leverages historical property trends, spatial amenity density, and price deviation metrics to identify undervalued or highgrowth-potential properties across urban areas. It combines location-sensitive scoring models with price per square foot analysis and Z-score based outlier detection to recommend listings that deviate positively from local price norms while offering strong amenity access. By evaluating properties based on proximity to hospitals, schools, banks, parks, transit, and other infrastructure, the model delivers context-aware investment insights. Key findings show the proposed model achieves a 70% match accuracy with expert evaluations, significantly outperforming baseline models. The implications of this work include a new framework for data-driven decision-making that can improve market efficiency, particularly in fragmented real estate markets like those in South Asia.