MPD: A Meteorological and Pollution Dataset: A Comprehensive Study of Machine and Deep Learning Methods for Air Pollution Forecasting

Alejandra Abalo-García, Sergio Hernández-García, Iván Ramírez, Emanuele Schiavi

Abstract: Air pollution is a significant global issue, being one of the leading causes of chronic diseases affecting the respiratory and neurological systems, and resulting in millions of deaths each year. Additionally, the scarcity of air quality sensors, due to their high cost, limits the availability of accurate data. In this study, we present a dataset that combines air quality and meteorological variables, with data sourced from the historical records of the Community of Madrid. Furthermore, we propose several baseline methods for this dataset. We then validate these baseline methods using another reference dataset, outperforming previous state-of-the-art methods. All the code and data is available in https://github.com/capo-urjc/MPD.git.