A data analytics approach for anticipating congested days at the São Paulo International Airport

Worldwide, most of the airports are not able to operate as planned due to delay problems. Since a high proportion of flights are affected by delays in congested days, for developing effective strategies to reduce flight delays and support response planning, a critical issue is how to anticipate the occurrence of congested days. The goal of this work is to employ a data analytics approach to build an early warning model to anticipate the occurrence of such days at the São Paulo International Airport. Therefore, a Mixture-of-experts model (MEM) was used to combine modelling approaches that rely on different assumptions regarding the data available to process. Such approach allows generating a more flexible and powerful model that makes good promises of improvement in the prediction accuracy. The built MEM is composed of a Classification and Regression Tree, a multiple linear regression and a seasonal ARIMA and it was used to generate predictions for three periods ahead. The accuracy of the early warning model was considered satisfactory for anticipating congested days.