Urban densities foster shorter network lengths

One argument for containing urban densities is that cities need a critical population density to sustain sufficiently available public transportation. However, the question of whether denser cities foster shorter public transport networks empirically is problematic, because real-world transport nets are a product of many additional factors that are presumably not related to urban form. I have recently published a paper in which I adopted a network expansion simulation approach to generate and analyze counterfactual data on network lengths for 36 world cities. To do so all networks are generated with similar expansion restrictions and objectives. Denser cities are found to have shorter simulated public transport networks, regardless of the tested model parameters. This provides additional proof that densities are needed to facilitate the provision of proximate public transport infrastructure, with potentially self-reinforcing effects.

The paper is published here. The GeoDMS scripts to generate counterfactual networks can be found on my github.

Transport Link Scanner

It has repeatedly been noted that there is a clear logic behind the geographical expansion of transport networks. That implies that one can model how those transport networks expand geographically; and how the final network would change if the network is constructed with different objectives.

In the last years I have developed the so-called Transport Link Scanner model, which is a GeoDMS based tool that allows the exploration of the effects of economic context and policy preferences on transport network expansion. The model combines a conditional logit model, some heuristics and techniques from the literature on corridor location problems and transport modelling methods to simulate the most likely network after the introduction of a new transport innovation.

The most recent version of Transport Link Scanner can be downloaded through this page. The scripts are downloadable here. The data is available here. Instructions to get you started are available here.

Converting public transport data into matrices

The recent years have seen a rise of websites that serve easily available public transport information for travellers; a great example is 9292 in the Netherlands. There is a substantial collection of data behind such websites, with many useful analytical and commercial possibilities. Those data are gradually becoming publicly available; see initiatives such as Transport for London’s.

For a scientific test case I recently developed a prototype method to generate network data and travel time matrices for the Netherlands using GeoDMS and a large raw database dump from the 9292 website. The generated networks have been combined with street data from OpenStreetMap to get a complete picture of travel time, including the necessary fore and after transport.

Some exciting things are possible with this sort of data. One application is generating accessibility maps and exploring how much public-transport enabled accessibility changes through the day. The top map show potential accessibility values by public transport for the Netherlands when arriving at five or six in the morning. The maps below display traveltimes to Amsterdam when leaving at different hours of the morning. These maps  clearly show the variability of access throughout the day – from some areas in The Netherlands it’s not even possible to reach Amsterdam early in the morning! Quite a shock for a country that claims to strive for a 24/7 economy!

Travel time to Amsterdam by public transport when leaving at the indicated time.
Travel time to Amsterdam by public transport when leaving at the indicated time.

One of my goals for 2016 is to build a program to generate spatiotemporally informed public transport networks, traveltime matrices and assorted network information from generic GTFS formatted files. When ready the methods will be made available through this website.

Using big data to understand mixed use benefits

In the 1960s Jane Jacobs shook urban planning with her famous work ‘The Death and Life of Great American Cities’. If you haven’t read it, give it a try, it’s a must-read! One of the things Jane Jacobs claimed after observing streetlife in her own neighbourhood was that city streets need to be occupied constantly to warrant safety and a pleasant atmosphere in public spaces. She claimed that mixed land-use patterns are necessary to provide constant occupation, and thus that there is a direct link between a city’s physical composition and its social sustainability. Unfortunately, apart from anecdotical evidence, proof from practice for Jacobs’s hypotheses has long been unavailable.

To verify whether land-use mixing indeed affects activity patterns I have recently used a massive dataset of mobile phone usage per antenna in Amsterdam, together with Piet Rietveld, Eric Koomen and Emmanouil Tranos. The used data are recorded by  Dutch mobile phone provider KPN and have kindly been provided by the Dutch ministry of Transportation.

That research proved beyond doubt that, yes, in Amsterdam mixed land-use patterns have more diverse and longer during activity patterns; and that in fact neighbourhoods that have more mixed land uses coincide with neighbourhoods that are more attractive. Thus these findings are, as far as I know, the first ever city-wide proof of Jane Jacobs’s assumptions. A full description of the methods and the results of this analysis have been published in Environment and Planning A. A summary of the research has been published in Rooilijn, a Dutch-language magazine for planning professionals.

For the Environment and Planning A article: click here.

For the Rooilijn article: click here.

 

GeoDMS learning

GeoDMS (‘Geo Data Model Server’) is an open-source software developed by ObjectVision that is very well-suited for managing and modelling large quantities of data. They are using it themselves for some pretty cool projects, see the gallery on the link above. It is the technical basis of large land-use modelling projects such as Land-Use Scanner and LUISA. Since 2008 I have used GeoDMS almost exclusively for the modelling work I have been involved in. I have used it myself to calculate travel times (here and here), compute accessibility values, generate public transport travel time matrices, develop transport models and simulate transport network expansion. If you are interested, GeoDMS can be downloaded via this link.

GeoDMS is an incredibly powerful tool, but even for those that have some experience in GIS programming it can be quite challlenging to start using it. The guys from ObjectVision have written an extensive manual for the software right here. While trying to teach the basics of the software, I found that the software needs a simple step-by-step explanation to get users started. Although I am not a big fan of ESRI’s ArcGIS suite, I have always admired their teaching material: that usually consists of very clear, step-by-step exercises that help make students feel at ease with the software. The GeoDMS community could surely use similar teaching material. For that reason I am currently writing an extended manual for GeoDMS. I am going to use this part of the website to add information, sample code and instructive movies for all you eager GeoDMS students.

 

How problematic is congestion? Travel times throughout the day

Congestion due to traffic has received a lot of political attention in the last decades. Costs are often mentioned as a reason why congestion should not be acceptable; while economists often uphold that, to some degree, congestion is a sign of economic success rather than a sign that a system is unsuccessful. In fact there is very little evidence that congestion on roads really hinders a city’s economic growth and that congestion only becomes a problem if traffic is severe throughout the day. If congestion is only a peak-time problem, spreading activities over the day becomes a feasible strategy to reduce travel delays.

Together with Paul Koster (VU University Amsterdam) I have edited data and made the video I share here. It shows real average working day travel times to Schiphol airport, the Netherlands for every quarter hour of the day. For Paul’s research and this video, the Dutch ministry of transportation kindly shared data on measured travel times for all of the motorways in the Netherlands.

The video shows that congestion is, from a time-of-day point of view, actually a very limited problem: before 7:00, between 9:00 and 16:00 and after 19:00 one can use the Dutch motorways presumably without any hindrance.

The effect of railroad network expansion on travel times in the Netherlands

The construction of overland transport infrastructure has had a major social and economic impacts by reducing travel costs and travel times. One of the key questions I am interested in is how travel time improvements have changed the level of accessibility and interaction opportunity within territories. I have made this movie to demonstrate the extent in which historical railroad construction in the Netherlands has affected travel times to the country’s capital, Amsterdam. Greater travel time improvements are demonstrated by lighter hues. The actual year is shown in the upper left corner.

To me, the movie and underlying data show two important things: first, that railroad construction has substantially altered travel times; second, that the largest improvements in travel times were realized in the first stages of railroad network development. The latest stages, when numerous local lines were added to the network, have much less effect on travel time savings.

Introduction

Hi, welcome! I am Chris Jacobs-Crisioni, the owner of Bureau Jacobs-Crisioni and this website. I consider myself an expert on developing and using support systems that help explore, understand and decide in questions related to sustainable urban planning and sustainable transport systems. I am experienced in using GIS software such as ArcGIS and in particular GeoDMS. This website is developed with three aims in mind: first, to introduce me and my work; second, to share some of the articles and code that I have written; and third, to have a platform to share training material for using GeoDMS software.