WasteWise Live Analytics

                   

About this Project: The Problem & Solution

                   

WasteWise Live Analytics solves the problem of measurable methane emissions by creating a reliable, low-cost system for real-time waste stream quantification. We successfully deployed a full-stack AI pipeline to replace unreliable manual auditing with instant, actionable data for robotic sorting. The primary output is item count, which eliminates the complex hardware dependency of visual mass estimation.

The system provides quantifiable environmental ROI C02 avoided) alongside operational data, which is crucial for facilities aiming to justify investment in automation and maximize the diversion of high-impact organic waste.

                                        

Key Features

                   

                       
• Live Item Counting via Instance Segmentation: Uses a trained YOLOv8 model to obtain pixel masks for accurate counting of items (Plastic, Metal, Organic, Glass) per frame.
                       
• Robust Cloud Bridge: Solved the complex network gap using ngrok tunneling to feed the private IP camera stream to the public Colab VM.
                       
• Bias Mitigation & Failover: Includes a software hack to redistribute the model's organic bias and a guaranteed mock output to ensure the live dashboard never fails.
                   

                   

Technologies Used

                   

• Machine Learning: YOLOv8n-seg (Instance Segmentation), PyTorch
• Data & Cloud: Firebase Firestore (Live DB), Google Colab (VM), `pyngrok` (Tunneling)
• Frontend & Viz: HTML, CSS, JavaScript (ES Modules), Chart.js, GitHub Pages (Hosting)

               

// Python fix to guarantee Firebase data integrity
upload_data = {
    'timestamp': firestore.SERVER_TIMESTAMP,
    'total_items_count': float(total_items), 
    
    // Explicitly cast Python Counter values to float for DB
    'item_distribution_count': { 
        cls: float(count) for cls, count in aggregated_item_counts.items()
    },
    
    'total_mass_kg_MOCK': float(total_mass), 
}