A New Timetable: from Rough Estimates to Fact-driven Insights from Smartcard Data
Niek Guis, Nederlandse Spoorwegen, Jan Banninga, Nederlandse Spoorwegen, Paul Siderius, ProRail
The introduction of a nationwide public transport smartcard system now provides a unique opportunity to analyse, explain and predict customer needs and travel behaviour better than ever before.
For public transport operators the line design and timetables are the backbone of the company strategies and policy. It determines the competitiveness over other modes, it dictates rolling stock allocation and thus buying strategies, it creates a demand for infrastructure (tracks, switches, platforms, stairs, etc) and a lot of other secondary services, like shopping facilities or bike-sharing programs follow from it. Optimising line design and timetables is thus one of the most important tasks for all public transport operators. This comes with careful thought and thorough knowledge of customer needs. Does a passenger prefer a long commuter train ride or will he wait for a faster Intercity alternative? Does he prefer a direct train with a low service interval, or would he prefer transferring with high frequent trains? Is he willing to pay an extra surcharge for a faster train or would he prefer the slower but cheaper train?
In the Netherlands the nation-wide introduction of smartcard travel was a true game changer in this field. All passengers have to check-in and check-out for each public transport trip. Billions of revealed travel transactions now provide a unique opportunity to analyse, explain and predict customer needs and travel behaviour better than ever before. For example, this has contributed to improved understanding of different travel behaviour for different trip purposes. This could eventually lead to adapted timetables for peak and off-peak.
The big dataset of train trip transactions was also used to recalibrate the transport models of NS (Netherlands Railways) and ProRail (infrastructure provider). Combined with the known demand patterns from smartcard data it is now possible to predict individual train loads throughout the day, with boarding and alighting during each train trip from start to end. It is also possible to predict these aspects for a complete new line design or timetable. In the past it was only possible to give a rough estimation on average working days or average peak hours based on counting and surveys. From now on, new line design scenarios or timetables can be evaluated extensively and in large detail. This provides unique insights, used in the iterative timetable optimisation process.
In the last year the new model was implemented and used in major policy decisions. Rolling stock allocation has been adjusted in order to better match travel demand and upcoming timetables have been created and refined thanks to the new model.
For instance, together with the redesign of the Amsterdam-Eindhoven corridor with high frequent trains, the service frequency of low-demand trains was evaluated by analysing travel demand for individual trains. This has led to extra trains on some routes on Sunday evenings and less trains on thin legs in the early weekend mornings. Policy decisions shifted from gut-feeling to fact driven insights based on smartcard data.
Association for European Transport