khushal-grover2005/LCA_predictor

🌍 AI-Driven Life Cycle Assessment (LCA) & Circularity Predictor

This repository houses the Inference Engine for an AI-powered platform designed to predict the environmental impact and circularity of metallurgical processes. Developed as part of a 14-day sprint, this system utilizes machine learning to transform complex industrial data into actionable sustainability insights.

🚀 Key Features

• —Dual-Target Prediction: Predicts Global Warming Potential (GWP) and Circularity Index simultaneously.
• —Sankey Diagram Logic: Generates source-target-value JSON flows to visualize carbon footprints from extraction to finished products.
• —Automated ML Pipeline: Includes modular components for Data Ingestion, Transformation (OHE, Log Scaling, Outlier Handling), and Model Training.
• —RESTful API: Flask-based backend ready for integration with React/Next.js frontends.

🛠️ Tech Stack

• —Language: Python 3.9
• —ML Frameworks: Scikit-Learn, XGBoost, CatBoost
• —API: Flask & Flask-CORS
• —DevOps: Docker, Hugging Face Spaces
• —Data Handling: Pandas, NumPy, Dill

📊 Project Structure

LCAPredictor/
├── artifacts/           # Saved model pickles and preprocessor
├── data/                # Raw and processed datasets
├── src/
│   ├── components/      # Data Ingestion, Transformation, Model Trainer
│   ├── pipeline/        # Training and Prediction Pipelines
│   ├── logger.py        # Custom logging module
│   └── exception.py     # Custom exception handling
├── app.py               # Flask API entry point
├── Dockerfile           # Container configuration
└── requirements.txt     # Project dependencies