This paper examines the effect of the COVID-19 induced lockdown upon six pollutants, CO, NO, NO2, PM10, PM2.5, and O3, in the Spanish community of Madrid. The paper relies on clustering methods and multiple regression techniques to control for a battery of potential confounding factors. The results show that the nationwide lockdown, decreed on 13 March by the Spanish government, exerted a statistically significant effect upon most pollution indicators. The estimates range from approximately −82% (NO and NO2) to −3% (CO). Reversely, the COVID-19 induced lockdown raised O3 levels by an average of 20%. By using data from 43 stations spread out among the region, the paper provides a local level analysis. This analysis reveals substantial differences across areas and across pollutants. This observation indicates that any successful approach to improve air quality in the region must be multidimensional.