Clustering Mobility-as-a-Service (MaaS) users through Gaussian Mixture Modelling and identifying factors affecting users' mode choice preferences

The concept of Mobility-as-a-Service (MaaS) has gained much popularity in the recent years to overcome issues pertinent to conventional transport systems, specifically in car-dependent societies to shift the travel behaviour towards more sustainable options. Despite numerous MaaS trials and implementations, existing MaaS studies mostly focus on the potential adoption and uptake of MaaS rather than analysing the actual MaaS users and understanding the characteristics of various users and their needs for more inclusive transport planning. Understanding the socio-demographics, travel resources and travel behaviour of MaaS users is important to evaluate the reach of MaaS and create strategies to enhance uptake among less-engaged populations. To address this gap, a revealed preference data of 2,182 respondents was collected through Breeze MaaS app and a cluster analysis approach for MaaS users based on the Gaussian Mixture Modelling (GMM) was proposed. After implementing GMM on the collected data from the Breeze MaaS app users, seven clusters were identified based on the mode share of participants in the Solent area of the UK. Based on the collected data, it was found that most of the MaaS users are young people, living mostly in urban areas and have more sustainable mode selection. Additionally, a Multinomial Logistic Regression (MNL) model was developed to comprehend the factors affecting the selection of different modes for each identified cluster compared to the car dependent users. The identified clusters together with the MNL model provide insights that could help a thorough understanding of actual MaaS users and guide targeted recommendations to increase engagement among current users. The findings could be used to reach a wider audience and increase the uptake of MaaS and sustainable mobility options in car-dependent areas.