🌿 GreenCO2

AI-Powered CO₂ Emission Monitoring & Compliance Platform

Python 3.11Flask 3.xReact 19PostgreSQL 15Docker Ready

Overview

GreenCO2 is a full-stack SaaS platform that helps companies track, analyze, and reduce their CO₂ emissions. It provides real-time dashboards with daily/monthly emission metrics, AI-powered predictions using Facebook Prophet for 7–14 day forecasts, and anomaly detection using Isolation Forest to flag unusual emission spikes.

It also features automated alerts delivered via email when thresholds are breached, and generates PDF compliance reports formatted against CPCB standards.

Features

🔐 JWT Auth: Stateless Bearer token authentication
📊 Dashboard: Live KPIs: today total, monthly total, 7-day trend
📝 Emission Logging: Multi-source entry (diesel, petrol, natural gas, electricity)
🔮 ML Predictions: Prophet-based 7–14 day CO₂ forecasts per company
🔬 Anomaly Detection: Isolation Forest flags unusual emission events
🚨 Smart Alerts: Threshold / trend / anomaly / prediction alerts
📧 Email Notifications: Multi-recipient HTML email alerts via SMTP
📄 PDF Reports: Professional compliance reports with ReportLab
⏰ Auto-Scheduler: Retrains models daily, sweeps alerts hourly

Architecture & ML Engine

The architecture consists of a React SPA browser frontend, communicating via HTTP/REST + JWT Bearer tokens to a Flask Backend running on Gunicorn. Data is persisted in PostgreSQL.

Browser (React SPA) 
      │ HTTP/REST + JWT Bearer 
      ▼ 
Flask Backend (Gunicorn) 
      │ 
      ├── Routes       → thin request/response handlers 
      ├── Services     → scheduler, business logic 
      ├── ML Engine    → Prophet + Isolation Forest models 
      │ 
      ▼ 
PostgreSQL Database

ML Scheduled Jobs

Every 24 hours: `train_all_models()` retrains Prophet for all companies.
Every 1 hour: Alert engine runs for all companies, emails new alerts.