🚗 Car Price Prediction API

A production-style Machine Learning API that predicts the selling price of a used car based on its specifications. The project demonstrates an end-to-end machine learning workflow, including data preprocessing, model training, model serialization, and deployment using FastAPI.

---

📌 Overview

Estimating the market value of a used car is an important problem for buyers, sellers, and dealerships. This project uses a machine learning model trained on historical car data to predict the expected selling price from various vehicle attributes.

The project is designed following a modular architecture to separate data preprocessing, model training, prediction, and API services, making it easy to maintain and extend.

---

✨ Features

• End-to-end Machine Learning pipeline
• FastAPI REST API for predictions
• Modular project structure
• Trained model serialization using Pickle
• Automatic API documentation with Swagger UI
• Data preprocessing pipeline
• Ready for cloud deployment

---

🛠 Tech Stack

• Python
• Scikit-learn
• Pandas
• NumPy
• FastAPI
• Uvicorn
• Pickle

---

📂 Project Structure

car-price-api/
│
├── app/
│   ├── main.py                 # FastAPI application
│   ├── schemas.py              # Request/Response models
│   └── routes.py               # API routes
│
├── src/
│   ├── data_preprocessing.py
│   ├── train.py
│   ├── predict.py
│   └── evaluate.py
│
├── models/
│   └── car_price_pipeline.pkl
│
├── data/
│   ├── train.csv
│   └── test.csv
│
├── run_training.py
├── requirements.txt
├── Procfile
└── README.md

---

🧠 Machine Learning Pipeline

The project follows the complete ML workflow:

1. Data Collection
2. Data Cleaning
3. Feature Engineering
4. Data Preprocessing
5. Model Training
6. Model Evaluation
7. Model Serialization
8. API Deployment

---

📊 Input Features

Example input:

{
  "year": 2018,
  "present_price": 7.5,
  "kms_driven": 25000,
  "fuel_type": "Petrol",
  "seller_type": "Dealer",
  "transmission": "Manual",
  "owner": 0
}

---

📤 Example Prediction

{
  "predicted_price": 6.82
}

---

🚀 Installation

Clone the repository

git clone https://github.com/Abhhiiissshhek/car_price_api.git

cd car_price_api

Create a virtual environment

python -m venv .venv

Activate it

Windows

.venv\Scripts\activate

Linux/macOS

source .venv/bin/activate

Install dependencies

pip install -r requirements.txt

---

▶️ Train the Model

python run_training.py

This will:

• preprocess the dataset
• train the regression model
• save the trained pipeline

---

🌐 Run the API

Start the FastAPI server

uvicorn app.main:app --reload

The API will be available at

http://127.0.0.1:8000

---

📚 Interactive API Documentation

FastAPI automatically generates API documentation.

Swagger UI

http://127.0.0.1:8000/docs

ReDoc

http://127.0.0.1:8000/redoc

---

🔌 API Endpoint

Predict Car Price

POST

/predict

Request

{
  "year": 2018,
  "present_price": 7.5,
  "kms_driven": 25000,
  "fuel_type": "Petrol",
  "seller_type": "Dealer",
  "transmission": "Manual",
  "owner": 0
}

Response

{
  "predicted_price": 6.82
}

---

📈 Model Evaluation

The model is evaluated using standard regression metrics such as:

• Mean Absolute Error (MAE)
• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)
• R² Score

---

☁️ Deployment

The project is designed for deployment on platforms such as:

• Render
• Railway
• Azure App Service
• AWS
• Docker

---

🔮 Future Improvements

• Docker containerization
• CI/CD with GitHub Actions
• Model versioning
• Authentication for API endpoints
• Request logging and monitoring
• Batch prediction endpoint
• Cloud deployment
• Interactive frontend using Streamlit

---

📚 What I Learned

This project helped me gain practical experience with:

• Machine Learning pipelines
• Feature engineering
• Model serialization
• REST API development using FastAPI
• Production-ready project structure
• Deploying ML models as web services

---

👨‍💻 Author

Abhishek Prajapati

• GitHub: https://github.com/Abhhiiissshhek
• LinkedIn: https://www.linkedin.com/in/abhishekprajapati-ml

---

⭐ Support

If you found this project useful, consider giving it a ⭐ Star on GitHub. Contributions, feedback, and suggestions are always welcome.