Modos de movilidad de los niños y niñas en edad escolar: exploración de la incidencia de factores socioeconómicos, de percepción y de mesoescala urbana utilizando Random Forest

Abstract

Analyzing and understanding the incidence of socioeconomic, perception and urban mesoscale factors on mobility modes of school-age children is essential to motivate a more sustainable mobility. This study explores the relationship between socioeconomic, perception and urban mesoscale factors on the daily mobility modes of school-age boys and girls (6 to 12 years old) in the intermediatesized city of Cuenca (Ecuador, Latin America). Random Forest, as a classifcation machine learning method, was used to classify the mobility modes into walk, bus and car, and to identify factor importance in each mode. The data were obtained from a mobility survey carried out on Cuenca households in 2019. Even if school bus is a usual mobility mode for schoolchildren, it could not be accounted in this study because it was not contained in the original survey. The results showed that the best model for walk and bus mobility modes was with all the factor groups, while for Car, as expected, was the socioeconomic model. Even if the most important factors were cars´ number per family and socioeconomic level, we also found that, in order to encourage walking as the mobility mode, the perception factors are relevant. Similarly, in order to encourage bus mobility mode, the urban mesoscale factors should be accounted for. This study contributes with data and a methodological approach that could infuence public policy regarding scholar-aged active mobility.