akeoj/Nigeria-real-estate-price-prediction

# Nigeria Real Estate Price Prediction ## Overview Real estate pricing is influenced by multiple economic, geographic, and property-related factors. This project explores the use of machine learning regression models to predict property prices in Nigeria based on structured features. The focus is on **model comparison, feature importance, and error analysis**, not just prediction. --- ## Problem Statement Accurate property valuation is critical for buyers, sellers, and investors. This project aims to estimate real estate prices using historical listing data and evaluate how different features contribute to pricing. --- ## Dataset - Domain: Nigerian real estate listings - Target variable: Property price - Features may include: - Location - Property type - Number of rooms - Property size - Dataset limitations such as missing values and regional bias are addressed during preprocessing. --- ## Methodology 1. **Data Preparation** - Data cleaning and normalization - Handling missing and inconsistent entries - Encoding categorical features 2. **Exploratory Analysis** - Price distribution analysis - Feature correlations - Regional price variations 3. **Modeling** - Baseline regression (Linear Regression) - Tree-based regression models - Comparison of model performance 4. **Evaluation Metrics** - Mean Absolute Error (MAE) - Root Mean Squared Error (RMSE) - Model generalization assessment --- ## Results - Tree-based models captured non-linear relationships better than linear baselines - Location and property size emerged as dominant price drivers - Error analysis highlighted areas where predictions are less reliable --- ## Interpretation & Use Cases Potential applications include: - Preliminary property valuation - Market trend analysis - Investment screening support Predictions should be combined with expert judgment and current market context. --- ## Limitations & Future Work - Limited geographic coverage may introduce bias - Market dynamics change over ti …