Raw data collected from NO2, O 3 and NO air pollution electrochemical low-cost sensors

Recently, the monitoring of air pollution by means of low-cost sensors has become a growing research field due to the study of techniques based on machine learning to im-prove the sensors' data quality. For this purpose, sensors un-dergo a calibration process, where these are placed in-situ nearby a regulatory reference station. The data set explained in this paper contains data from two self-built low-cost air pollution nodes deployed for four months, from January 16, 2021 to May 15, 2021, at an official air quality reference sta-tion in Barcelona, Spain. The goal of the deployment was to have five electrochemical sensors at a high sampling rate of 0.5 Hz; two NO2 sensors, two O 3 sensors, and one NO sensor. It should be noted that the reference stations publish air pol-lution data every hour, thus at a rate of 2 . 7 x 10 -4 Hz. In ad-dition, the nodes have also captured temperature and relative humidity data, which are typically used as correctors in the calibration of low-cost sensors. The availability of the sen-sors' time series at this high resolution is important in order to be able to carry out analysis from the signal processing perspective, allowing the study of sensor sampling strategies, sensor signal filtering, and the calibration of low-cost sensors among others. (c) 2022 The Authors. Published by Elsevier Inc.This is an open access article under the CC BY-NC-ND license( http://creativecommons.org/licenses/by-nc-nd/4.0/ )

Automated summary

Recently, there has been increasing research on monitoring air pollution using low-cost sensors and improving sensor data quality through machine learning techniques. This paper presents a dataset from two self-built low-cost air pollution nodes deployed at an official air quality reference station in Barcelona, Spain. The dataset includes four months of data from five electrochemical sensors as well as temperature and relative humidity data. The availability of high-resolution sensor time series is crucial for analyzing sensor sampling strategies, signal filtering, and calibration of low-cost sensors.