Recovering information from synthetic air quality index

Synthetic indices are often used to condense complex situations into a single figure. However, this condensing process risks losing potentially useful information, especially when the index is to be utilised by public decisionmaking bodies. The present study proposes a general strategy, combining a number of different methods, designed to recover information from air-quality indices: graphical methods to reconstruct the composition of pollution, multinomial logit analysis to study the influence of meteorological covariates on air-quality indices, and finally, a probability distribution for the index itself as a basic tool with which to interpret the index’s crucial values.