Shaping the Transit Landscape of the Netherlands with Conveyal Analysis

Andrew Byrd
Conveyal
Published in
5 min readJan 17, 2017

Conveyal has partnered with Dutch transportation engineering firm Movares to assess the impact of long-term public transport development scenarios for the Netherlands using our next-generation Conveyal Analysis platform. After a pilot project in Rotterdam and The Hague (the southern wing of the heavily urbanized, polycentric Randstad region), our collaborative accessibility-based methods were adopted by the Dutch Ministry of Infrastructure and the Environment and applied in all regions of the country. We recently completed the process for Metropolitan Amsterdam, and the resulting regional development priorities will be integrated into the national Public Transport 2040 Future Vision.

Innovative aspects of Conveyal Analysis that made it a great fit for this study include:

  • A focus on accessibility indicators (who can reach what in how much time) rather than mobility or infrastructure-based indicators (traffic speed or volume)
  • Automated creation of a baseline multi-modal network model from up-to-date public data sources in open standard formats
  • Interactive, map-based creation of future scenarios, with changes influencing results in a matter of seconds
  • Highly optimized and parallelized calculation of accessibility indicators, enabling quick turnaround and collaborative exploration of numerous scenarios’ impact at locations throughout a metropolitan region
  • Rigorous consideration of variations in travel time and accessibility capturing the full range of passenger experiences at different departure times. This includes waiting time uncertainty on hypothetical future routes where only headway is specified rather than full schedules.

Context: The Netherlands

The Netherlands has a long history of accessibility-based spatial planning. It is a fitting place to apply our next-generation methods and software, as many of us at Conveyal were first exposed to accessibility concepts via Dutch planning policy over the past several decades. Of course sophisticated traffic models already exist in the Netherlands, but at the time of this study they had not been updated to include certain essential public transportation services, and the difficulty and expense of doing so was prohibitive. Perhaps more importantly, existing models had not been designed to allow collaborative creation of scenarios and rapid, interactive visualization of their implications for access to employment and other opportunities.

Building a Multi-modal Network Model

A bicycle park-and-ride garage with train departure boards in Delft, southern Randstad

Using two widely available and frequently updated data sources, Conveyal Analysis can build a complete, up-to-date model of the entire Dutch street and transit network. When corrections are necessary, the network can be updated and rebuilt in a matter of minutes. The transit portion is built from schedule data in the GTFS format, which is published by transit operators and authorities for use in trip planners such as Google Maps or OpenTripPlanner. The Netherlands has perhaps the highest quality and most complete open GTFS data in the world, with schedules for every bus, train, tram, and ferry in the country (as well as real-time position and delay updates for every vehicle). The street network is built from OpenStreetMap data, which contains information about the kinds of vehicles that can use each street, the direction of traffic flow, sidewalks, turn restrictions, park-and-ride lots, etc. This global data source can be audited, extended and updated by its users. Dutch OpenStreetMap coverage is excellent, including detailed bicycle and pedestrian paths which Conveyal Analysis takes into account in all travel time calculations.

Adding Demographic Data

In Conveyal Analysis, demographic and opportunity data are represented on detailed regular grids, eliminating the distortions and inaccuracies that result when populations are grouped into large zones such as postal codes or census tracts. For Metropolitan Amsterdam, we built these density grids from population and job counts at the individual building level from the LISA database. Future land use was projected based on output from the Amsterdam VENOM model and the Netherlands Regional Model.

Projected spatial distribution of jobs (blue) and housing (red) in Metropolitan Amsterdam derived from building-level data

Collaborative Scenario Creation and Analysis

Modifying the transit network in the Conveyal Analysis web-based scenario editor

Many potential modifications to Amsterdam’s regional transit system were evaluated separately before being integrated into complete scenarios. Accessibility indicators were calculated for points throughout the region under micro-scenarios that added or removed single routes or tested individual changes to frequencies, speeds, and route alignments. Informed by Conveyal Analysis visualizations and maps, planners and stakeholders considered network effects while iteratively assembling these measures into packages. The final scenario that emerged from discussions was generalized into broader statements about which investments the region would like to make, accompanied by cost estimates. These regional conclusions are now being integrated into a nationwide plan that determines which projects will be realized and how funding will be allocated.

Notable Outcomes

Change in the average number of jobs accessible from each point in Metropolitan Amsterdam under a complex scenario

Conveyal Analysis helped planners make better-informed decisions about which services to retain or eliminate. Our accessibility indicators, which are sensitive to transfer time, showed that more frequent services requiring transfers were a suitable alternative to lower-frequency direct buses into rural areas.

Our results also strongly supported a plan to allocate separate tracks to light rail and intercity traffic in the Schiphol airport tunnel. Local heavy rail service could be replaced by extending Amsterdam metro and tram lines through the tunnel. This strategy was initially proposed as a way to reduce congestion, but now enjoys broad support since Conveyal Analysis revealed an unexpected benefit: it increases accessibility to the job market for outlying villages by eliminating transfers.

Transforming Transportation and Land Use Planning

Using open data, Conveyal and Movares were able to build a Dutch network model that was more current and detailed than existing transport models, even reflecting transfer timings and barriers to pedestrian or bicycle movement such as waterways. Our innovative algorithms allowed us to apply speculative headway-based scenarios, maintaining detailed schedules on the unmodified part of the baseline network and characterizing any uncertainty and variability in results. Planners were able to visualize and discuss the accessibility impact of each proposed change individually before assembling them into a final vision. With Conveyal Analysis, planners in the Netherlands can now benefit from continuous, immediate, and nuanced feedback while regional transportation scenarios are still being sketched out. Don’t hesitate to contact us if you’d like to discuss bringing similar capabilities to your region.

--

--