MRC Fly Management System
A robust fly management system providing real-time tracking, data visualisation, and user-friendly interfaces for efficient fly control and monitoring across multiple locations.
- >Real-time sensor data ingestion pipelines
- >Recharts data visualization interfaces
- >Role-based control panels with audit logging
Project Vision & Meaning
The MRC Fly Management System is an industrial IoT solution designed for facilities requiring strict sanitation compliance (food processing factories, warehouses, pharmaceutical labs). Historically, monitoring pest traps required daily manual logs by inspection teams. MRC Fly digitizes the process by integrating smart sensors into fly traps, enabling automated detection, counting, and instant alerts for safety supervisors.
Real-World Problems Solved
It eliminates manual inspection overhead by automating trap audits, ensures safety standard compliance (HACCP/FDA) through audit trails, and triggers instant alerts during count spikes to prevent major pest infestations.
Challenges & Solutions
Telemetry Ingestion Spikes and Database Overload
Problem Context:With hundreds of smart sensors sending activity pings simultaneously, direct database writes caused connection exhaustion and led to intermittent telemetry data loss during peak activity hours.
Created an ingestion buffer pipeline on the Node.js API server. Incoming pings are gathered in memory and written to PostgreSQL in batches every 5 seconds. Implemented an exponential backoff retry mechanism inside the IoT clients to handle network dropouts without losing count data.
Visualizing Large Pest History Datasets Without Lag
Problem Context:Recruiters and auditors review multi-month trap data across multiple locations. Fetching and rendering these large datasets directly in standard browser-based chart libraries (like Recharts) resulted in browser UI freezes and long page loads.
Implemented the Largest Triangle Three Buckets (LTTB) downsampling algorithm on the backend to reduce 100,000+ data points to 500 visual-representative points before transmitting them. This preserved visual trends while reducing data payload size by 99.5%, bringing chart rendering speeds under 100ms.