🏠 Housing Price Prediction

This project aims to predict housing prices using machine learning models and analyze the importance of housing features. The dataset contains structured information about real estate properties including area, number of rooms, location factors, and furnishing status.

---

📁 Project Structure

housing-price-prediction/
│
├── data/
│   ├── Housing.csv                  # Original dataset
│   └── housing_cleaned.csv          # Cleaned dataset with additional features
│
├── notebooks/
│   ├── 01_data_preparation.ipynb    # Data cleaning, encoding, outlier detection
│   └── 02_model_training.ipynb      # Model training, evaluation, feature importance
│
├── requirements.txt                 # Required Python libraries
└── README.md                        

📊 Dataset Overview

The dataset includes:

• area, bedrooms, bathrooms, stories
• mainroad, guestroom, basement, etc.
• furnishingstatus: encoded in multiple formats (ordinal and one-hot)
• Engineered features:
  • area_per_bedroom
  • area_per_bathroom
  • is_fully_equipped

---

Workflow Summary

1. Data Cleaning & Encoding

• Converted object columns to numeric/boolean
• Used one-hot and ordinal encoding strategies
• Detected and removed outliers using Isolation Forest
• Created engineered features

2. Model Training

Trained and evaluated the following models:

Model	Encoding Used
Linear Regression	One-Hot
Ridge Regression	One-Hot
Random Forest Regressor	Ordinal
Gradient Boosting Regressor	Ordinal
XGBoost Regressor	Ordinal

All models were compared using:

• RMSE
• MAE
• R² Score
• Actual vs Predicted scatter plots

3. Feature Importance

• Coefficients (linear models)
• Feature importances (tree models)

---

Next Steps

• Fine-tune hyperparameters
• Use SHAP values for explainability
• Explore ensemble and stacking models
• Deploy a trained model as an API

---