Publicação

Predicting Station Occupancy on Bike-Sharing System During Events: The Lisbon Case Study

Ver documento

Detalhes bibliográficos
Resumo:The Internet of Things (IoT) uses intelligent technology and interconnected devices to gather real-time data, enabling a shift to sustainable transportation modes. This has made bikesharing systems (BSS) a popular and reliable form of soft mobility in urban environments. This study leverages these technological advancements to understand how city events affect Lisbon's BSS, GIRA, by predicting hourly station occupancy rates using the CRISP-DM methodology. By integrating weather information and event data into BSS station-level data spanning the year of 2022, and applying state-of-art machine learning (ML) algorithms – such as Random Forest (RF), Gradient Boosting Tree (GBT), and Extreme Gradient Boosting (XGBoost) – the research aims to optimize station occupancy management during events. This contributes to more efficient urban transportation systems and a sustainable future for Lisbon. Major findings show that XGBoost outperformed the other algorithms, having a higher predictive accuracy during sport event days, and in music event days stations near Coliseu dos Recreios require strong rebalancing efforts.
Autores principais:Oliveira, Rita Madalena Cardoso da Silva
Assunto:Bike-sharing Soft Mobility Machine Learning Predictive Modeling Cycling SDG 11 - Sustainable cities and communities SDG 13 - Climate action
Ano:2024
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:The Internet of Things (IoT) uses intelligent technology and interconnected devices to gather real-time data, enabling a shift to sustainable transportation modes. This has made bikesharing systems (BSS) a popular and reliable form of soft mobility in urban environments. This study leverages these technological advancements to understand how city events affect Lisbon's BSS, GIRA, by predicting hourly station occupancy rates using the CRISP-DM methodology. By integrating weather information and event data into BSS station-level data spanning the year of 2022, and applying state-of-art machine learning (ML) algorithms – such as Random Forest (RF), Gradient Boosting Tree (GBT), and Extreme Gradient Boosting (XGBoost) – the research aims to optimize station occupancy management during events. This contributes to more efficient urban transportation systems and a sustainable future for Lisbon. Major findings show that XGBoost outperformed the other algorithms, having a higher predictive accuracy during sport event days, and in music event days stations near Coliseu dos Recreios require strong rebalancing efforts.