A decade of air quality in Bogotá: a descriptive analysis

In this work we apply a rigorous and reproducible data analytics process for validation and analysis of the historical data from Bogotá (Colombia) air quality monitoring network since 1998. The reasons for addressing this research study stem from the lack of a consistent approach for cleaning, validating and reporting air quality data. By analyzing the whole dataset, we are aiming at providing citizens and the city authorities with a clear view of the current situation of air quality and of its historical evolution. Without any loss of generality, we focus our analysis on both respirable and fine particulate matter (PM10 and PM2.5) concentrations, which in Bogotá and worldwide are source of concern for their negative impacts on human health. We develop a reproducible and flexible data cleaning methodology for particulate matter concentration data reported by the local authorities, which allows customizing and applying configurable validation rules. Then, we present statistical descriptive analyses by providing intuitive data visualizations, characterizing historical and spatial change of air pollutant levels. Results raise concerns for the high percentage of invalid data, as well as the high levels of PM2.5 and PM10 ambient concentrations as observed in the valid portion of the available data, which frequently exceed national and international air quality standards. The data exhibit encouraging signs of air quality improvement, particularly for PM10. However, the analyses indicate that significant differences exist across Bogotá, and particularly in the south-west zone of the city annual concentrations of particulate matter are up to three or four times the WHO recommendations. We are confident on the methodology and results from our analysis are useful both for local environmental authority and the general public to help in obtaining consistent conclusions from the available data