Device health monitoring system project for AirQo

Abstract

AirQo is an initiative deploying low-cost sensors across Africa to provide real-time air quality data. This project addresses critical operational challenges faced by AirQo's hardware team, who currently employ a reactive approach to device management, resulting in extended periods of compromised data quality and inefficient resource allocation. We developed a comprehensive Device Health Monitoring System to enable proactive maintenance and data-driven decision making. The system employs a three-layer architecture: a Data Sources Layer collecting telemetry data from GSM-enabled sensors; a Backend Infrastructure Layer processing this data using Python and Apache Airflow for orchestration with PostgreSQL for storage; and a Frontend Layer built with Next.js for visualization. Key functionalities include automated anomaly detection, performance tracking, and predictive maintenance scheduling. The system calculates critical metrics such as Mean Time Between Failures (MTBF) and Mean Time To Recovery (MTTR) while analyzing performance patterns across deployment locations. Implementation utilized containerized deployment with Docker to ensure environment consistency and independent scaling. The technology stack was selected based on operational requirements: Thing Speak for IoT data handling, Python and Airflow for data transformation, PostgreSQL for time-series optimization, FastAPI for API services, and Next.js for frontend performance. Results demonstrate significantly improved monitoring capabilities, with the system detecting potential failures before they impact data quality and optimizing maintenance resource allocation. This enhanced pipeline establishes a foundation for scalable monitoring as AirQo's sensor network expands, ultimately contributing to improved air quality data reliability and public health outcomes across Africa.