Full-stack Machine Learning system for used car price prediction in Morocco – from data scraping to model deployment via FastAPI, integrated into a web application.
# AutoPredict
AutoPredict is a used car price prediction project for the Moroccan market.
The project covers the complete pipeline: data collection, machine learning and deep learning modeling, deployment through a REST API, and integration into a web application.
## Technologies Used
**Data & ML:** Python, Pandas, Selenium, BeautifulSoup, scikit-learn, TensorFlow
**Backend:** FastAPI, Node.js, Express, MongoDB
**Frontend:** HTML, CSS, JavaScript
**Security:** JWT authentication
## Branch Structure
**master:** Web interface and deployment (Node.js + MongoDB + frontend)
**ml_pipeline:** Jupyter notebooks containing all ML stages:
- Data scraping and cleaning
- Dataset preparation
- Training and comparison of models (Linear Regression, Polynomial Regression, Decision Tree, Random Forest, SVR, ANN)
- Model testing and validation
## Key Features
- Data collection and CSV dataset preparation
- Training and evaluation of regression models and neural networks
- Model deployment via REST API using FastAPI
- Full-stack web interface for accessing predictions
- JWT authentication, prediction storage, and history tracking