Replication data for "Deep Learning for Citywide Air Quality Inference: Cross-location Forecasting from Sparse Stations"

This dataset contains historical air quality measurements from two major urban areas: Madrid, Spain and Cali, Colombia, obtained from api.aqi.in. The data were collected from 15 fixed monitoring stations in Madrid and 18 in Cali, offering a comprehensive view of air pollution levels in each city over time.

Each record includes pollutant concentrations captured at 5-minute intervals. The pollutants and their corresponding measurement units are as follows:

Additionally, it contains the following meteorological data:

Finally, each measurement contains a topic value, an integer that serves to uniquely identify the monitoring station that produced it, and the timestamp.

Data Collection and Processing:

1. Data was retrieved via the endpoint: https://api.aqi.in/api/v1/getMonitorsByCity, with the corresponding city specified in the request header.
2. Records were collected at 5-minute intervals and stored in a MongoDB database in JSON format whenever new data became available.
3. Non-essential fields (e.g., units, location name) were removed.
4. Sensor topics were simplified by retaining only the final four digits of the topic tag.
5. In Cali, 39 sensors that appeared only once were excluded to ensure data reliability.

Source of Data:

This dataset contains air quality data originally retrieved via the api.aqi.in platform. The data is sourced from public air quality monitoring stations operated by:

• Alcaldía de Santiago de Cali for data related to Cali, Colombia (https://www.cali.gov.co/dagma/publicaciones/38365/sistema-de-vigilancia-de-calidad-del-aire-de-cali-svcac/)

• Ayuntamiento de Madrid for data related to Madrid, Spain (https://airedemadrid.madrid.es/portal/site/calidadaire)

Rights and Usage:

The original data is made available through public monitoring systems by the respective municipal governments listed above. The data was accessed via api.aqi.in, which aggregates publicly available air quality information. This dataset is shared for academic and research purposes only, and to the best of our knowledge, the underlying data is in the public domain.

Funding:

This work was supported by the grant PID2023-148214OB-C21 funded by MICIU/AEI/10.13039/501100011033 and by FEDER/EU. This work was also supported in part by the grants TED2021-129336B-I00 funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR. This work was also supported by the grant PCI2024-153485 funded by MICIU/AEI/10.13039/501100011033 and by the European Union. This research was also funded by the PRIMA Programme under Grant Agreement No. 2431 (FUSION: Comprehensive and sustainable solution to minimize food loss and waste and promote food security in the Mediterranean region). This work was also funded by Fundación Séneca (22236/PDC/23). This work was also a result of the ThinkInAzul and AgroAlNext programmes, funded by Ministerio de Ciencia, Innovación y Universidades (MICIU) with funding from European Union NextGenerationEU/PRTR-C17.I1 and by Fundación Séneca with funding from Comunidad Autónoma Región de Murcia (CARM). The work of Francisco-Jose Alvarado-Alcon was supported by the Spain’s Ministry of Universities under Grant FPU22/00316.