EcoPackAI
An intelligent system that leverages machine learning to predict packaging costs and CO₂ environmental impact, helping businesses make sustainable packaging decisions backed by data.
Technologies Used
Problem
Packaging decisions in supply chains are often made without considering environmental impact. There was no accessible tool that could predict both cost and sustainability metrics simultaneously.
What I Built
A full-stack application with a trained ML model that takes packaging parameters as input and predicts both the cost and carbon footprint, displayed through an interactive dashboard with exportable reports.
Challenges
Training accurate predictive models with limited data, building a robust feature engineering pipeline, and creating an intuitive dashboard that makes complex ML predictions accessible to non-technical users.
Architecture
Flask REST API serving a scikit-learn ML model, with PostgreSQL for data persistence and SQLAlchemy as the ORM. JavaScript frontend consuming the API and rendering interactive data visualizations.
Key Features
- ML prediction pipeline for cost and CO₂ impact estimation
- Feature engineering with pandas and NumPy
- Interactive analytics dashboard
- Sustainability reports generation
- PDF and Excel export capabilities
- Flask REST API backend
- PostgreSQL database with SQLAlchemy ORM