Integrating Predictive Modelling and Low-Cost Sensor for Air Quality Management and Public Health Intervention in Semi-Urban Nigeria.

Abstract

Air pollution driven by rising atmospheric CO₂ concentrations represent a global health issue across the globe, which majorly affecting developing nations based on limited or no monitoring infrastructure, where planning intervention and effective policy are essential through use of predictive models that integrate real-time data with computational approaches. This study addresses the critical gap in air quality monitoring for semi-urban industrial communities in Nigeria by developing predictive models using low-cost sensor data. The study measured the in-situ data of atmospheric CO₂ concentrations, temperature, relative humidity, and pressure over three small-scale industries (Garri processing factory, bakery, and Ebu) in Nigeria, using PCE-AQD 50 sensor. Daily data were modified from 30 seconds data for modelling using Python programming language. OLS regression and SARIMAX models were employed to predict air quality trends over the study locations, while R², RMSE, and MAPE metrics assessed the model performance. The study developed models where SARIMAX model demonstrated superior predictive performance with R² = 0.89, and RMSE = 0.15, as well as OLS Regression revealed significant correlations between CO₂ concentrations and meteorological parameters. It was equally found that predictive simulations indicated that without intervention, CO₂ concentration levels in industrial communities could exceed WHO-recommended thresholds by 2030, while under moderate emission control scenarios, reduction of up to 25% is achievable. Hence, the research provides the first comprehensive prediction framework for air quality trends in semi-urban Nigerian industries, offering evidence-based insights for policymakers, enabling proactive public health interventions that safeguard public health and climate change mitigations, and demonstrating the potential of low-cost sensor to enhance air quality management in resource-constrained environments.