Smart Mobility Improvement: Classifying Commuter Satisfaction in Sydney, Australia

Abstract

This paper attempts to derive useful insights from commuter feedback data. It investigates transportation mode, commuting density and peak hours in Sydney, Australia. Machine Learning techniques are then applied to analyse traveler satisfaction to discover useful models for classification. Experiments demonstrate that each method has its competitive advantages over others, and no approach completely outperform other methods in terms of accuracy, performance, and interpretability. It is suggested that one could use Support Vector Machine to classify satisfied commuters, and/or utilize Neural Network to classify unsatisfied travelers.