Exploring the access/egress behaviour of rapid rail passengers through the application of nested logit models

Abstract

The Gautrain is a rapid rail system in Gauteng, a province of South Africa. Currently, little is known about what causes the behaviour of Gautrain passengers concerning their choice of mode for their first- and last-mile trips. This thesis is a study of the first- and last-mile mode choice behaviour of Gautrain passengers. The study had three main aims. First, it aimed to develop the most accurate and statistically significant models of both first- and last-mile behaviour. Part of this aim was to try to understand the effect of non-traders on the stated preference data. To achieve this end, models were estimated with and without non-traders in the sample. Second, the study attempted to determine if there were any differences between commuters’ first- and last-mile behaviour. For this purpose, a stated preference survey was developed and put online. It was then marketed on the social media platforms of the Gautrain Management Agency, the body that manages the rail system. The results of the survey were analysed and cleaned of errors. Different discrete choice models were derived from the data to find the best fitting model structure. The last aim was to see how the developed models could be used to plan future access and egress services. Significant discrete choice models were developed on a reasonably representative sample of the Gautrain population. The best-fitting model structure was a nested one for both first- and last-mile trips. These models show that while first- and last-mile commuter behaviour was similar, there were some differences. In particular, people are much more sensitive to invehicle time for their last-mile journey in comparison to their first-mile journey. This exercise showed that although non-trader data reduced the effectiveness and significance of the model slightly, the data did not change the overall picture shown by the models without non-traders.