IoT-Based Air Quality Monitoring System for Urban Environmental Awareness: A Deployment and Validation Study in Jakarta and Semarang

Abstract

Rapid urbanization in Indonesia has intensified air pollution in major cities, even as monitoring infrastructure remains geographically sparse. This paper presents the development and field validation of a low-cost, community-deployable IoT-based air quality monitoring system built to close that coverage gap. The system integrates four sensors GP2Y1010AU0F (PM2.5), MQ-7 (CO), MQ-135 (CO₂), and DHT11 (temperature/humidity) with an ESP32 microcontroller, transmitting readings every four seconds to a cloud backend that serves both a web dashboard and a mobile app. Two nodes were deployed in Palmerah (West Jakarta) and Banyumanik (Semarang), with validation readings collected on the same calendar day (August 1, 2024) at both sites and benchmarked against 20 paired hourly AQI readings per site (n=40) from the Nafas Indonesia commercial platform. The West Jakarta node achieved a mean absolute error (MAE) of 10.40 AQI units (RMSE=11.97, R2=0.63), while the Semarang node achieved an MAE of 22.50 units (RMSE=25.07, R2=0.20); both sites showed a consistent negative bias, with the system underestimating relative to the commercial reference. Despite this, the system correctly classified the dominant air quality risk category in all 40 comparison observations. These single-day results indicate the system is a promising tool for community-level air quality awareness, and demonstrate a reproducible, sub-USD-30 architecture for expanding environmental monitoring coverage in support of SDG 3 and SDG 11 pending multi-day validation before broader claims of field reliability are warranted.