Data-driven public transport ridership prediction approach including comfort aspects

Abstract

The most important aspects on which passengers base their choice whether to travel by public transport are the perceived travel time, costs, reliability and comfort. Despite its importance, comfort is often not explicitly considered when predicting demand for public transport. In this paper, we include comfort level in a modelling framework by incorporating capacity in the public transport assignment. This modelling framework is applied in the public transport model of HTM, the urban public transport operator of The Hague. The current transportation demand is directly derived from smart card data and future demand is estimated using an elasticity based approach. The case study results indicate that not considering capacity and comfort effects can lead to a substantial underestimation of effects of certain measures aiming to improve public transport (up to 30%). We also illustrate that this extended modelling framework can be applied in practice: it has a short computation time and leads to better predictions of public transport demand.