Exploring the influence of socio-economic aspects on the use of electric scooters using machine learning applications: A case study in the city of Palermo

Abstract

Most European countries have been committed to reducing their carbon footprint, combating climate change, and reducing the air pollution typical in large cities over the past decade. Among current solutions that can be adopted are the replacement of fuel-powered means of transport with electric ones, as well as the introduction of car sharing, bike sharing and electric scooters.

The post-pandemic phase was characterized by a greater propensity to use these means of transport as they were perceived as a healthier choice (for a greater possibility of implementing social distancing) and cheaper (for the diffusion of shared services). The study of modal choice depends on socio-economic structures. The present work analyses data related to socio-economic factors (work, income and other) to examine the tendency to use electric scooters in the metropolis of Palermo, Sicily, through machine learning algorithms.

The comparison of different algorithms allowed us to underline how the multilayer perceptron algorithm obtained the best classification among the minimal sequential optimization algorithms. The findings also highlight middle-income and freelancer people as being more likely to use micro-mobility than others. Contrary to what was thought, these findings revealed that micro-mobility is not just a preferred mode of transport for low-income people or students. These trends will be able to encourage continuous monitoring of the relevant factors and will be able to help political decision-makers to increase and improve the diffusion of micro-mobility and to direct marketing campaigns to the groups identified here.