Overview

Linear-Lasso-Regression-Car-Price-Prediction is an interactive web application that predicts the price of a car (or house) using machine learning models: Linear Regression and Lasso Regression. Built with Streamlit, it provides an intuitive interface for users to input car features and instantly get price predictions using two different regression models.

• Author: Nhan Pham
• Email: ptnhanit230104@gmail.com
• Version: 1.0.0
• Created: 2025-07-26

Visual Workflow

flowchart TD
    A[User Inputs Car Features] --> B{Choose Model}
    B -- Linear Regression --> C[Preprocess Features]
    B -- Lasso Regression --> C
    C --> D[Load Model]
    D --> E[Predict Price]
    E --> F[Display Result in UI]

Features

• User-friendly Web UI: Built with Streamlit for easy interaction.
• Dual Model Support: Predict prices using both Linear Regression and Lasso Regression.
• Instant Results: Get predictions in real-time as you input data.
• Educational: Explore and compare the effects of different regression techniques.
• Extensible: Modular codebase for easy extension to other regression models or datasets.
• Open Source: Freely available for learning, research, and extension.
• Well-documented: Includes guides for dataset, models, and Streamlit usage.
• Actively Maintained: Issues and PRs are welcome!

Table of Contents

• Overview
• Visual Workflow
• Features
• Demo
• How It Works
• Example Prediction
• Technical Stack
• Dataset
• Model Training
• Installation
• Usage
• Advanced Usage
• Project Structure
• Extending the Project
• Troubleshooting & FAQ
• Contributing
• Changelog
• Learning Resources
• Requirements
• License
• Acknowledgements
• Contact

How It Works

1. User Input: Enter car features (year, price, kms driven, fuel type, seller type, transmission, owner).
2. Model Selection: Choose between Linear Regression or Lasso Regression.
3. Preprocessing: The app encodes categorical features and prepares the input for the model.
4. Prediction: The app loads the selected model (from .pkl files) and predicts the price.
5. Result Display: The predicted price is shown instantly in the UI.
6. Model Comparison: Users can easily compare predictions from both models to understand the impact of regularization.

Example Prediction

Suppose you want to predict the price of a car with the following features:

• Step 1: Enter these values in the app fields.
• Step 2: Click With Linear model or With Lasso model.
• Step 3: The app will display something like:

Predicted Price: $420,000.00

Try changing the model or input values to see how the prediction changes!

Technical Stack

• Frontend/UI: Streamlit (Python-based web app framework)
• Backend/ML: scikit-learn (for model training and inference)
• Data Handling: pandas
• Visualization: matplotlib (for EDA and model training)
• Serialization: pickle (for saving/loading models)
• Jupyter Notebook: For model development and experimentation

Dataset

• Source: See dataset/car_data.csv
• Description: Contains 301 rows and 9 columns, including features like Year, Present_Price, Kms_Driven, Fuel_Type, Seller_Type, Transmission, Owner, and the target Selling_Price.
• Details:
  • No missing values
  • Categorical features are encoded for model compatibility
  • For more, see docs/dataset.md
• Sample Columns:
  • Year: Year of manufacture
  • Present_Price: Current ex-showroom price
  • Kms_Driven: Kilometers driven
  • Fuel_Type: Petrol/Diesel/CNG
  • Seller_Type: Dealer/Individual
  • Transmission: Manual/Automatic
  • Owner: Number of previous owners
  • Selling_Price: Price at which the car was sold (target)

Model Training

• Notebook: model/model-training.ipynb
• Models:
  • linear_model.pkl: Trained Linear Regression model
  • lasso_model.pkl: Trained Lasso Regression model
• Workflow:
  1. Data loaded and cleaned
  2. Categorical features encoded
  3. Data split into train/test sets
  4. Models trained and evaluated
  5. Best models saved as .pkl files
• Learn more:
  • Linear Regression Guide
  • Lasso Regression Guide

Installation

1. Clone the Repository

git clone https://github.com/NhanPhamThanh-IT/Linear-Lasso-Regression-Car-Price-Prediction.git
cd Linear-Lasso-Regression-Car-Price-Prediction

2. Install Dependencies

Create a virtual environment (optional but recommended):

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

Install required packages:

pip install streamlit scikit-learn pandas matplotlib

3. (Optional) Retrain Models

If you want to retrain the models, open and run model/model-training.ipynb in Jupyter Notebook or VSCode.

Usage

1. Run the App

streamlit run app/main.py

The app will open in your browser at http://localhost:8501.

2. Using the App

• Fill in the car features in the input fields
• Click With Linear model or With Lasso model to get a prediction
• The predicted price will be displayed instantly
• Try different inputs and compare model results

Advanced Usage

• Custom Port: Run on a different port:

  streamlit run app/main.py --server.port 8502

• Headless Mode: For deployment:

  streamlit run app/main.py --server.headless true

• Docker: See docs/streamlit.md for Docker deployment instructions.
• Streamlit Cloud: Deploy directly from GitHub for free.
• API Integration: (Advanced) Wrap prediction logic in a FastAPI or Flask API for programmatic access.

Project Structure

Linear-Lasso-Regression-Car-Price-Prediction/
├── app/
│   ├── main.py           # Streamlit app entry point
│   ├── predictor.py      # Model loading and prediction logic
│   └── ui.py             # Streamlit UI components
├── dataset/
│   └── car_data.csv      # Car price dataset
├── docs/
│   ├── dataset.md        # Dataset documentation
│   ├── lasso-regression-model.md  # Lasso regression theory & practice
│   ├── linear-regression-model.md # Linear regression theory & practice
│   └── streamlit.md      # Streamlit learning guide
├── model/
│   ├── lasso_model.pkl   # Trained Lasso model
│   ├── linear_model.pkl  # Trained Linear model
│   └── model-training.ipynb # Model training notebook
├── LICENSE
└── README.md

Extending the Project

Want to add new features or support more models? Here are some ideas:

• Add More Regression Models: Integrate Ridge, ElasticNet, or custom models.
• Feature Engineering: Add new input features or transformations.
• Visualization: Show feature importance, residuals, or model diagnostics in the UI.
• API Integration: Expose predictions via a REST API for integration with other apps.
• Deployment: Deploy to Streamlit Cloud, Heroku, or Docker (see docs/streamlit.md).
• UI Enhancements: Add charts, explanations, or user authentication.

How to add a new model:

1. Train and save your model as a .pkl file.
2. Update app/predictor.py to load and use the new model.
3. Add a button or option in app/ui.py for users to select the new model.
4. Update the README and docs as needed.

Troubleshooting & FAQ

Q: The app doesn't start or crashes on launch.

• Make sure all dependencies are installed (pip install streamlit scikit-learn pandas matplotlib).
• Check your Python version (should be 3.7+).
• Ensure you are running the command from the project root directory.

Q: I get a ModuleNotFoundError for 'streamlit' or 'sklearn'.

• Activate your virtual environment if you created one.
• Run pip install -r requirements.txt if you have a requirements file.

Q: The prediction is always the same or seems off.

• Check that the input features are reasonable and within the expected range.
• Make sure the model .pkl files are present in the model/ directory.
• Retrain the models if needed using the provided notebook.

Q: How do I deploy this app online?

• See deployment instructions in docs/streamlit.md for Streamlit Cloud, Heroku, and Docker.

Q: Can I use my own dataset?

• Yes! Replace dataset/car_data.csv with your own data (matching the expected columns), retrain the models, and update the code as needed.

Contributing

Contributions are welcome! To contribute:

1. Fork the repository
2. Create a new branch (git checkout -b feature/your-feature)
3. Make your changes and commit them
4. Push to your fork (git push origin feature/your-feature)
5. Open a Pull Request describing your changes

Guidelines:

• Write clear, descriptive commit messages
• Document new features or changes in the README/docs
• Follow PEP8 style for Python code
• Add tests or example usage if possible
• Please read our Code of Conduct before contributing.

Changelog

See CHANGELOG.md for a list of major changes, new features, and bug fixes.

Learning Resources

• Streamlit Guide
• Linear Regression Guide
• Lasso Regression Guide
• Dataset Guide
• scikit-learn Documentation
• Streamlit Documentation
• Pandas Documentation

Requirements

• Python 3.7+
• streamlit
• scikit-learn
• pandas
• matplotlib

You can also create a requirements.txt file for easy installation:

streamlit
scikit-learn
pandas
matplotlib

License

This project is licensed under the MIT License.

Acknowledgements

• Streamlit
• scikit-learn
• pandas
• matplotlib
• UCI Machine Learning Repository (for dataset inspiration)

Contact

For questions, suggestions, or contributions, please contact:

• Nhan Pham
• Email: ptnhanit230104@gmail.com