EcoPackAI

1. Project Title

EcoPackAI - AI Framework for Sustainable Packaging Recommendation

2. Short Project Description

EcoPackAI predicts packaging cost and CO2 impact for packaging materials and ranks them using an eco score to recommend sustainable options.

3. Milestone Status (Current)

Milestone 4 completed with:

• working ML pipeline + backend API
• integrated frontend recommendation studio
• BI dashboard with CO2 reduction, cost savings, and usage trend charts
• sustainability report export (PDF + Excel)
• deployment-ready configs for Render/Heroku

4. Tech Stack

Layer	Technology
Language	Python
API	Flask
ML/Data	scikit-learn, pandas, numpy, joblib
DB (optional scripts)	PostgreSQL
DB Access	SQLAlchemy, psycopg2-binary
Frontend	HTML, CSS, JavaScript

5. Installation

1. Clone:

git clone <your-repo-url>
cd EcoPackAI

2. Create venv:

python -m venv .venv
# PowerShell
.\.venv\Scripts\Activate.ps1

3. Install dependencies:

pip install -r requirements.txt

4. Configure environment:

copy .env.example .env
# then update DB_PASSWORD (and other values if needed)

6. Project Structure (Important Files)

backend/
  app.py                        # Main API used by frontend
frontend/
  index.html                    # Dashboard UI
  script.js                     # API integration + rendering
  style.css                     # UI styling
ml/
  generate_realistic_dataset.py # 600+ realistic row generation
  data_quality_gate.py          # Quality validation checks
  dataset_preparation.py        # Feature prep + scaler + splits
  model_training.py             # Cost/CO2 model training
  recommendation_engine.py      # CLI recommendation script
  predict_api.py                # Legacy/simple ML API
  dataset/
    real_materials_base.csv
    real_production_materials.csv
    materials.csv
    cleaned_materials.csv
    engineered_materials.csv
  models/
    cost_model.pkl
    co2_model.pkl
    scaler.pkl
dashboard/
  analytics.py                  # Metrics + top material analysis
report/
  EcoPackAI_Milestone3_Presentation.pptx

7. End-to-End Run Order (Current)

Run in this order:

python ml/generate_realistic_dataset.py
python ml/data_quality_gate.py
python ml/dataset_preparation.py
python ml/model_training.py
python backend/app.py

Open frontend:

• frontend/index.html in browser
• API base URL: http://127.0.0.1:5000

8. API Endpoints (backend/app.py)

• GET /health -> service status, dataset row count, active feature schema
• GET /metadata/material-types -> material type list for frontend dropdown
• GET /recommend -> best + top-5 recommendation from dataset
• POST /recommend -> recommendation from custom material input
• GET /analytics/summary -> BI dashboard metrics + chart data
• GET /reports/sustainability?format=pdf|excel -> downloadable sustainability report

Sample payload:

{
  "materials": [
    {
      "material_name": "Custom Candidate",
      "material_type": "Bioplastic",
      "strength_rating": 7.5,
      "weight_capacity": 40,
      "biodegradability_score": 8,
      "recyclability_percentage": 82
    }
  ]
}

9. What Is Done

• Realistic synthetic dataset generation from original base dataset
• Minimum 600+ rows achieved (currently 700+)
• Data quality checks (range, nulls, distribution, correlations)
• Train/test split, scaling, model artifact saving
• Leakage-prone feature usage reduced in core training/inference flow
• Backend API integrated with frontend
• Frontend renders recommendation cards/tables (not raw JSON dump)
• BI dashboard for analytics insights (CO2 reduction, cost savings, trends)
• PDF + Excel sustainability report export
• Milestone 3 PPT created in report/

10. Current Limitations

• No automated unit/integration test suite yet
• No Docker/CI pipeline yet
• Some legacy scripts are still present and not part of new primary flow
• DB scripts contain hardcoded credentials and need env-based cleanup
• backend/models/ has placeholder 0-byte model files; active models are in ml/models/

11. Recommended Next Steps

• Add run_pipeline.py to chain generate -> quality gate -> prepare -> train
• Add tests for API schema and ML pipeline contracts
• Move all secrets/config to .env
• Add Docker + CI workflow
• Add model versioning and retraining logs

12. Deployment (Render/Heroku)

The repo includes deployment-ready configuration:

• Procfile for Heroku (gunicorn backend.app:app)
• render.yaml for Render

Render

1. Create a new Web Service and connect the repo (or use Blueprint to auto-read render.yaml).
2. Build command: pip install -r requirements.txt
3. Start command: gunicorn backend.app:app --bind 0.0.0.0:$PORT
4. Add env vars from .env (DB_* if using PostgreSQL).

Heroku

1. heroku create
2. git push heroku main
3. heroku config:set your env vars (DB_* if using PostgreSQL)

13. License

No license file currently included. Add a LICENSE file (MIT/Apache-2.0 recommended).