kellyose/nigeria-houses-price-prediction

# nigeria-houses-price-prediction This project predicts house prices in Nigeria using Linear Regression and XGBoost models. The dataset includes features like bedrooms, bathrooms, toilets, parking spaces, house type, town, state, and price. The code preprocesses the data, trains the models, and evaluates their performance. ##Project Overview Goal: Predict house prices in Nigeria. Dataset: nigeria_houses_data.csv (not included; provide your own). Models: Linear Regression and XGBoost. Features: Numerical (e.g., bedrooms) and categorical (e.g., town, state). Preprocessing: Feature engineering, scaling, encoding, log transformation. Dataset Columns: bedrooms, bathrooms, toilets, parking_space, title, town, state, price. Size: 24,326 entries. Source: User-provided CSV (e.g., /content/drive/MyDrive/nigeria_houses_data.csv). Requirements Python 3.7+ Libraries: numpy, pandas, matplotlib, seaborn, scikit-learn, category_encoders, xgboost ##Results Linear Regression MAE (NGN): 0.43 (likely a bug; expected in millions). R² (log scale): 0.6500. R² (original scale): 0.6586. Top Features: town, state_Plateau. XGBoost MAE (NGN): 53,851,020.25. R² (log scale): 0.7136. R² (original scale): 0.7138. Top Features: title_Detached Duplex, town. XGBoost outperforms Linear Regression in R², but MAE suggests room for improvement.