Predicting House Prices in Abuja Using Machine Learning Models

Accurate house price prediction supports informed property investment decisions and real estate market transparency in Nigeria's fastest-growing property market. This study developed and compared machine learning models for predicting residential property prices in Abuja Municipal Area Council and Bwari Area Council, Federal Capital Territory. A dataset of 3,200 residential property listings was compiled from three online real estate platforms covering 2021 to 2023, including features for location, size, number of bedrooms, proximity to amenities, estate type, and construction quality rating. After preprocessing and feature engineering, linear regression, random forest, and XGBoost models were trained and tested using an 80-20 split. XGBoost achieved the lowest mean absolute error of N2.4 million on the test set and an R-squared of 0.87, outperforming the linear baseline by 31 percent. Location district and property size were the most important features in the XGBoost model. Properties in Maitama, Wuse 2, and Asokoro commanded prices 3.2 to 4.7 times higher than comparable properties in Lugbe and Nyanya. Properties marketed exclusively through social media were systematically undervalued by online platforms. The study contributes to transparency in a market historically dominated by informal valuations. Recommendations include development of a publicly accessible Abuja property price prediction tool, collaboration with REDAN on data standardisation, and longitudinal dataset expansion to cover outer satellite towns in the FCT.

Keywords: house price prediction, XGBoost, real estate analytics, Abuja, FCT