Low-Cost Smart Sensing Pipeline: Assembly, Calibration, and Interpretation of Air Quality Data

Monitoring air pollutant concentrations, particularly particulate matter (PM2.5 and PM10) and CO2, is critical for urban environmental management and public health. Traditional monitoring systems are often expensive, limiting their widespread use. This paper presents a comprehensive, low-cost air quality sensing pipeline that integrates hardware design, machine learning-based calibration, deployment, and spatial data visualization for decision-making. We designed and assembled a sensor using accessible components and developed a machine learning framework using Support Vector Regression and Random Forest to calibrate the sensor's measurements against referencegrade equipment. The pipeline was deployed in an urban area in Brazil, demonstrating a complete integration of the stepssensor assembly, calibration, deployment, and data visualization-within one system, providing a scalable, costeffective solution for air quality monitoring in smart cities.