Public Transport Trip Route Planner Utilizing Historic Delay and Crowding Data

Public Transport Trip Route Planner Utilizing Historic Delay and Crowding Data


Olav Kåre Malmin, SINTEF Technology and Society, Petter Arnesen, SINTEF Technology and Society, Erlend Dahl, SINTEF Technology and Society


This paper describes a project to use historical data for delays and crowding in the public transport system of Oslo, and develop a PT trip planner application using this data to minimize time and reduce risk of missed transit.


This paper describes one of the main activities in the Norwegian Stratmod project presented at ETC 2015 (Nordheim, Tørset: Developing a policy model for sustainable urban transport, 2015). The aim of this activity is to use collected real-time data about delays and crowding from the public transport operator Ruter AS in Oslo as input to the policy model.

The public transport operator in Oslo (Ruter AS) provided a log file containing all public transport activity in Oslo between March and April 2015. The most notable data fields described actual arrival/departure versus scheduled arrival/departure, and the number of passengers entering or exiting each vehicle at each bus stop. Based on this data average delays and crowding for each stop along each route for different times of the day were calculated.

The new Stratmod policy model utilize delays and crowding between OD-pairs, in addition to various time cost components and number of trips. The Norwegian Regional Transport Model calculates costs and passenger trips between OD-pairs, but does not calculate delays or crowding.
To find the delays and crowding between OD-pairs we proposed to develop a trip planner application for public transport that cannot only output delays and crowding data, but also use delays and crowding to find the best path between two points of choice, like the origin and destination points in the transport model. The main functionality in this trip planner compared to other trip planning applications is using average delays for both time minimization and finding transit points. Using delays to find transit points will reduce the risk of arriving too late for the next leg of the trip.

Since we have crowding data for various public transport lines for various times of day, we expand the trip planner to incorporate crowding and find routes trying to minimize the risk of crowding. This works in two dimensions: find a longer route that is less popular, or wait until the best route is less crowded.

The operator Ruter AS uses the open source OpenTripPlanner software ( for their online public trip planning service. We will further develop on the OpenTripPlanner software to get comparable results with the Ruter service. The main input to the trip planner is Ruter's transit lines on the GTFS data format openly available (


Association for European Transport