Air quality low-cost sensors and monitoring stations NO2 raw dataset in Rouen (France)
Data was collected by ATMO Normandie using 9 low-cost electrochemical sensors and 2 monitoring stations (thereafter named QDP and SUD3) located in Rouen, France, close to major roads. At the beginning of the measurements campaign, each sensor is installed co-located with one traffic monitoring stati...
Gespeichert in:
1. Verfasser: | |
---|---|
Format: | Dataset |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Data was collected by ATMO Normandie using 9 low-cost electrochemical sensors and 2 monitoring stations (thereafter named QDP and SUD3) located in Rouen, France, close to major roads. At the beginning of the measurements campaign, each sensor is installed co-located with one traffic monitoring station (4 at the SUD3 site and 5 at the QDP one). Most of them were permuted to the other one so as to study the stability of the calibration models when the device is turned off, moved and installed in a different location. The dataset gathers hourly averaged data given by low-cost sensors and monitoring stations between October 20, 2021 to March 25, 2022. It also includes co-variates measured by the sensors: temperature, relative humidity, atmospheric pressure, plus Ox and CO measures. The devices being co-located, the main goal is to calibrate low-cost sensors by means of a model which should provide a NO2 measurement in μg/m3. Then, the model is to be tested against another monitoring station by moving the sensors, so as to study its stability when the environment is modified. In this aim, some sensors were also moved a second time to make new measurements on the site they were calibrated. Moreover, this experiment was undertaken so as to understand how moving or shutting down the devices impacts the efficiency of calibration models. However, two sensors were never moved from their original location, so as to study the existence of a temporal drift. Our full dataset is divided into 9 CSV files (one per sensor) named ‘ASExx.csv’ where ‘xx’ denotes the sensor’s id. (‘xx’ goes from 4 to 13, excluding 9). Each file is organized as follows, having for columns:
- date: the timestamp in UTC, in the format yyyy-mm-dd HH:MM.
- ASE_NO2: the hourly averaged concentration of NO2 measured by the sensor in µV.
- ASE_NO: the hourly averaged concentration of NO in µV.
- ASE_CO: the hourly averaged concentration of CO in µV.
- ASE_Ox: the hourly averaged concentration of Ox in µV.
- ASE_T: the hourly averaged temperature inside the node in °C.
- ASE_HR: the hourly averaged relative humidity inside the node in %.
- ASE_PA: the hourly averaged atmospheric pressure in hPa.
- SUD3_NO2: the hourly averaged concentration of NO2 measured by the reference monitor SUD3 in μg/m3.
- SUD3_NO: the hourly averaged concentration of NO measured by the reference monitor SUD3 in μg/m3.
- QDP_NO: the hourly averaged concentration of NO measured by the reference monitor QDP in μg/m3.
- QDP_NO2: the |
---|---|
DOI: | 10.17632/82dnstrd93 |