Analysis of travel patterns from precarious settlements transit users in São Paulo through smart card data mining.

• Ano de defesa: 2018
• Autor(a) principal: Pieroni, Caio De Borthole Valente
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Clustering algorithms; Engenharia de transportes; Mineração de dados; Mobilidade urbana; Planejamento de transportes; Precarious settlements; Smart card data; Transportation planning; Transporte público; Travel patterns
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3138/tde-29032022-100650/

Smart Card Data (SCD) allow us to understand and to analyze mobility at an exceptional level of detail. However, they can be considered restricted when analyzing users trip purposes. Identifying travel patterns may provide better context to smart card data. More specifically, this identification may allow the understanding of travel patterns of transit users from precarious settlement areas, a portion of the population that historically has limited and unequal access to financial resources and opportunities. This work aims to understand the temporal and spatial patterns of urban transit movements of residents of precarious settlements in the city of São Paulo, through smart card data mining. For this, we apply three distinct clustering algorithms: K-means, TwoStep, and Self Organizing Maps (SOM). Residents of middle-class areas of the city are also included to compare the behavioral differences in urban displacements in the studied areas as a function of their residents household income. The results showed that the clusters formed by the three methods show similar results, and clusters with high number of commuters mostly composed by precarious settlement residents suggest an association of this residents with low-paid employment, with their smart card transactions, mainly registered in residential medium / high-income and residential low-income land use areas.