Leveraging Open Data and Land Use Regression Models for Scalable Air Quality Monitoring: Integrating Low-Cost Sensors for Global Applicability

The growing availability of open geospatial data presents significant opportunities to address spatial and temporal gaps in air quality (AQ) monitoring. This is particularly crucial in resource-limited and underserved regions. In this study, we introduce a novel framework that combines low-cost sensor (LCS) measurements with Land Use Regression (LUR) models, utilising publicly accessible datasets to improve the spatial and temporal resolution of AQ estimates. By integrating empirical sensor data with open, openly available model predictors, the framework enhances the accuracy of air quality predictions, particularly in areas with limited high-density monitoring infrastructure.

The open LUR model uses freely available data, such as traffic density, land cover, population distribution, and meteorological information, from OpenStreet Map and OpenWeatherMap, to predict local pollutant concentrations. These predictions are dynamically calibrated with real-time LCS measurements through machine learning regression techniques, which adjust for sensor biases, reduce noise, and quantify uncertainties. This integration allows for a more accurate, real-time representation of air pollution levels, especially in urban areas where traditional monitoring is often inadequate or nonexistent.

To demonstrate our framework's capability in refining Nitrogen Dioxide (NO₂) and Particulate Matter (PM₂.₅) estimates, we conduct a study in a dense population area in Nairobi, Kenya. Our result achieves a significant improvement in alignment with regulatory-grade measurements. By relying on open data and open-source LUR models, this approach is scalable, adaptable, and transferable, making it a cost-effective solution for AQ monitoring in diverse geographic and socio-economic settings.

This work emphasises the transformative potential of open data and open LUR models in democratising access to high-resolution, real-time air quality monitoring tools. By combining low-cost sensors with these open-source data and models, the framework offers actionable insights for urban planning, public health initiatives, and environmental policy. It underscores the broad applicability of this solution in addressing global air pollution challenges, providing scalable tools for effective air quality management and policy-making across various regions.

 

Keywords:

Low-cost sensors (LCS), open data, open Land Use Regression (LUR) models, air quality monitoring, Nitrogen Dioxide (NO₂), Particulate Matter (PM₂.₅), urban planning, public health, environmental policy, data fusion, scalability, global applicability, open-source models.