Development of a multiple regression model to calibrate a low-cost sensor considering reference measurements and meteorological parameters.

Low-cost air quality sensors are widely used to improve temporal and spatial resolution of air quality data. In Lima, Peru, only a limited number of reference air quality monitors have been installed, which has led to a lack of data for establishing environmental and health policies. Low-cost technology is promising for developing countries because it is small and inexpensive to operate and maintain. However, considerable work remains to be done to improve data quality. In this study, a low-cost sensor was installed with a reference monitor station as the first stage for the calibration process, and a multiple regression model was developed based on reference measurements as an outcome variable using sensor data, temperature, and relative humidity as the predictive parameters. The results show that this particular technology exhibits a promising performance in measuring PM2.5 and PM10 (particulate matter with diameter aerodynamic less than 2.5 μm and 10 μm, respectively); however, the correlation for PM2.5 appears to be better. Temperature and relative humidity data from the sensor were only partially analyzed due to the evident low correlation with the reference meteorological data. The objective of this study is to begin analyzing the performance of low-cost sensors that have already been introduced to the Peruvian market and selecting those that perform better to provide for informed decision-making.