A Low-Cost Distributed IoT Sensor Network for Real-Time Air Quality Management: A Case Study at Federal Polytechnic Oko, Nigeria

This study presents the design and deployment of a low-cost, distributed IoT sensor network for real-time air quality monitoring across Federal Polytechnic Oko, Nigeria. Using ESP32 microcontrollers integrated with SenseAir S8 (CO₂) and Plantower PMS5003 (PM2.5) sensors, the system generated high-resolution, spatially differentiated data over a 14-day period. Results revealed three critical hazards: (i) critical indoor ventilation failure in Lecture Classrooms, with CO₂ concentrations reaching 3,150 ppm; (ii) severe outdoor pollution at Generator Points, where PM2.5 peaked at 135.0 µg/m³ (a 540% exceedance of WHO guidelines); and (iii) occupational safety risks in Science Laboratories, with NO₂ spikes of 410 ppb. Comparative validation against a commercial IAQ meter confirmed that single-point monitoring underestimated CO₂ and PM2.5 peaks by up to 60% and failed to detect NO₂ entirely, underscoring the necessity of distributed sensing. To extend beyond passive monitoring, a Random Forest model was trained to forecast PM2.5 one hour ahead, achieving R² = 0.87 and 91% exceedance accuracy. SHAP analysis confirmed that recent PM2.5 values and generator-related NO₂ spikes were the strongest predictors, ensuring model transparency and stakeholder confidence. Together, the IoT network and AI forecasting framework provide actionable policy intelligence, enabling targeted interventions such as ventilation protocols and generator management strategies. This work establishes a scalable, cost-effective blueprint for sustainable environmental governance and enhanced public health protection in high-density academic environments across West Africa.