Predicting air pollution in Chile: A statistical modeling perspective

Carolina Marchant Fuentes

Abstract:

Rising levels of air pollution worldwide have led to a range of adverse effects on human health. According to a recent World Health Organization study, nine out of ten people globally breathe air with high pollutant levels, resulting in seven million deaths annually. This issue is particularly significant in several Chilean cities. The World Air Quality Index Ranking, which measures air quality based on fine particulate matter levels, places Chile second only to Peru in Latin America and the Caribbean for cities with the highest levels of fine particulate matter (https://www.iqair.com/world-air-quality-ranking). This evidence highlights a severe public health crisis in Chile, especially during the winter months when particulate matter concentrations peak. In this context, we present predictive models designed to estimate particulate matter levels based on climatological and meteorological variables. Specifically, we explore multivariate, semi-parametric, and machine learning models, applying them to real air pollution data from Chilean cities using the R-project software. Our results demonstrate that these models are effective in predicting extreme urban air pollution episodes, thereby enabling the prevention of adverse health effects on the Chilean population.

Keywords: air pollution; particulate matter; predictive model; semi-parametric models.

Link do seminário: https://teams.microsoft.com/meet/279371837520299?p=ooGZx6qF5OxJqXBGdQ