Car Price Prediction

A machine learning web application that predicts the selling price of used cars based on various features like age, mileage, fuel type, and more.

Features

• Predict car selling prices using a trained Random Forest model
• Clean, responsive web interface
• Support for multiple fuel types (Petrol, Diesel, CNG)
• Considers transmission type, seller type, and ownership history

Demo

Enter your car details and get an instant price prediction:

Input	Description
Year	Year the car was purchased
Showroom Price	Original price in Lakhs
Kilometers Driven	Total distance driven
Previous Owners	Number of previous owners (0, 1, or 3+)
Fuel Type	Petrol, Diesel, or CNG
Seller Type	Dealer or Individual
Transmission	Manual or Automatic

Installation

1. Clone the repository:

git clone https://github.com/RoshaniPawar16/Car-Price-Prediction.git
cd Car-Price-Prediction

2. Create a virtual environment (recommended):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Run the application:

python app.py

5. Open your browser and go to http://127.0.0.1:5000

Project Structure

Car-Price-Prediction/
├── app.py                              # Flask web application
├── main.py                             # Alternative entry point
├── templates/
│   └── index.html                      # Web interface template
├── random_forest_regression_model.pkl  # Trained ML model
├── car data.csv                        # Training dataset
├── Untitled.ipynb                      # Model training notebook
├── requirements.txt                    # Python dependencies
├── Procfile                            # Heroku deployment config
└── README.md

Model Details

• Algorithm: Random Forest Regressor
• Hyperparameter Tuning: RandomizedSearchCV with 5-fold cross-validation
• Best Parameters:
  • n_estimators: 1000
  • max_depth: 25
  • max_features: sqrt
  • min_samples_split: 2
  • min_samples_leaf: 1

Model Performance

Metric	Value
MAE	0.88 Lakhs
MSE	3.95
RMSE	1.99 Lakhs

Feature Importance

The most influential features for price prediction:

1. Present Price (37.6%)
2. Fuel Type - Diesel (22.8%)
3. Seller Type - Individual (13.4%)
4. Transmission - Manual (13.4%)
5. Car Age (8.1%)

Dataset

The model was trained on a dataset of 301 cars with the following features:

• Car Name, Year, Selling Price, Present Price
• Kilometers Driven, Fuel Type, Seller Type
• Transmission, Owner

Technologies Used

• Backend: Python, Flask
• ML: scikit-learn, pandas, numpy
• Frontend: HTML, CSS
• Deployment: Heroku (Gunicorn)

License

This project is open source and available under the MIT License.