Abstract

This study presents the application of self-organizing maps to air-quality data in order to analyze episodes of high pollution in Madrid. The goal of this work is to explore the dataset and then compare several scenarios with similar atmospheric conditions : some of them when no actions were taken and some when traffic restrictions were imposed. The levels of main pollutants, recorded at these stations for eleven days at four different times from 2015 to 2018, are analyzed in order to determine the effectiveness of the anti-pollution measures. The visualization of trajectories on the self-organizing map let us clearly see the evolution of pollution levels and consequently evaluate the effectiveness of the taken measures, after and during the protocol activation time.