Short term ridership prediction in public transport by processing smart card data
Public transport operators are exposed to massive data collection from their smart card systems. In the Netherlands, every passenger needs to check in and to check out, resulting in detailed information on the demand pattern. In buses and trams, checking in and checking out takes place in the vehicle, providing good information on route choice. This paper explores options for using this smart card data for analysis and performing what-if analyses by using transport planning software. This new generation of transport demand models, based on big data, is an addition to the existing range of transport demand models and approaches. The intention is to provide public transport operators with a simple (easy to build) model to perform these what-if analyses. The data is converted to passengers per line and an OD-matrix between stops. This matrix is assigned to the network to reproduce the measured passenger flows. After this step, what-if analysis becomes possible. With fixed demand, line changes can be investigated. With the introduction of an elastic demand model, changes in level of service realistically affect passenger numbers. This method was applied on a case study in The Hague. We imported the smart card data into a transport model and connected the data with the network. The tool turned out to be very valuable for the operator to gain insights into the effect of small changes.
Read the paper: TRB 2015
Comments are closed.