TACQ: How experts around the world are using new technology to assess cycle infrastructure

New to the #DCENetwork, Roxy Tacq, offers independent consultancy services in urban mobility, cycling safety, and road safety with her company TACQ. One of her skills is combining software with her mobility background to assess and improve cycling facilities. 

Using a program called CycleRAP, images of cycle paths around the world can be assessed by experts on a set of criteria to generate a safety score and report. We sat down with Roxy to learn more about the use of this tool and how it can shape the future of cycle mobility.  

Could you introduce yourself and the work that you do? 

I’m Roxy Tacq. I have been working as an independent advisor since the beginning of 2022. Before that, I worked at the ANWB for more than 12 years as a public affairs advisor. There, I gained a lot of insight into mobility and that’s also how I became acquainted with the iRAP program, which CycleRAP is a part of. The ANWB is one of the founding members of EuroRAP, which is now the global iRAP program, active in 180 countries.  

What is the goal of assessing roads with CycleRAP?  

The goal of the road assessment is to give people insights into safety of the infrastructure and how it can be improved, and to enable improvements in a cost-effective manner. Doing that we can avoid crashes and save lives. The clients are often a local, regional or national government who want to get those insights. The inspection and project management are usually done by an accredited supplier, like me.  

There are also organisations who want to use this as an advocacy tool. Because maybe their local government is not investing (enough) in cycling safety, the design of the facilities is poor. You can make a stronger case with a report based on an evidence-based model, which is used worldwide and says: “it is not safe enough and you need these investments to make it safer”. 

If someone reaches out to you and says that they want to assess their road network. What are the next steps?  

A project consists of taking images of the road, with geolocation. And then enter data based on those images. Once you have all this data, the model produces risk scores of all road sections and then it’s analysis and reporting. I’ve worked on CycleRAP projects in collaboration with two parties. Both with a different approach.  

One is with a Spanish company, Lane Patrol, who developed the software to do projects efficiently. They have an Android app which gives the option for organizations to capture images of the infrastructure themselves. Then the team in Spain will do the data analysis of the different road features because they’re trained to do that efficiently. Based on the data and CycleRAP risk scores I will make an analysis and report. This report gives insights based on the data on where the low and high-risk zones are and what kind of risk we are looking at. So that’s not only crashes with cars, but also with cyclists, single bicycle accidents, pedestrian accidents, and most important: how can we improve it?  

You can also do it on a network-level, like a whole city or a whole region. If there are countries who have a lot of data, like the Netherlands, then it might not make sense to collect all the data based on the images. You might know where the facilities are, the speed limits, if there’s lights along the road or not. If there is a lot of reliable data available, then we only have to collect a number of characteristics. Or we could give a concept score based on data of ~ 90% of the characteristics. With this approach we can provide a risk assessment and preliminary investment plan for every piece of infrastructure used by cyclists (paths and roads) in a whole province. This is something I offer with DOK data, a company from the Netherlands. 

So, it really depends on the client, how much data they have and the scope of their request. My role is to find the right solution and then deliver insights from the analysis. 

How can the program help bring unique projects to life? 

A good example is using CycleRAP in the design phase. If you have plans to implement something, you can already make calculations on the safety level of the facilities you’re going to design. You can then compare this design to alternatives (wider facilities, segregating them with barriers) and compare the safety impact. So you can include this in your cost-benefit analysis. The CycleRAP risk model has been developed in such a way that it’s suitable for any country. There are also some projects where they use machine learning to assess the images of facilities. There are some suppliers investing in that, but it also takes some time to develop it and you need a lot of images. Both Lanepatrol and DOK data are exploring these opportunities. 

At the same time, CycleRAP suppliers are using the conventional method of trained professionals who label all the images. That will help us to do it automatically later. But of course, we also need labeling of the images in different circumstances. I’ve done one of these projects to test if it could be done with the iRAP on the provincial roads in the Netherlands. With computer vision we could reliably code around 70% of the road attributes based on the images. However, you need images of the facilities from your region and also have sufficient images of special circumstances to be able to reliable detect them. 

Therefore, I urge governments not to wait for these technologies to be fully ready, but to already start using this method of assessments. Furthermore, through the iRAP program, we are already saving lives by recommending road safety interventions that offer strong cost-benefit outcomes. 

What differences do you see in cities outside the Netherlands? 

Some of the challenges we have here [in the Netherlands] are different from other cities. Like overcrowding of our facilities and a greater diversity of bicycle and micromobility on the facilities. But I think the struggles we have will come to other cities soon, if they have not already. The whole reason we started CycleRAP >10 years ago is that we foresaw that we were going to have more and more cyclists; and at the same time an aging population. This would mean more single and multiple bike crashes. I know that in many other countries cyclists are already happy if there is any type of bike lane available.  

Now, in cities like Barcelona, you see with the uptake of delivery by bike (on fast e-bikes) and people using e-scooters, it’s getting busier, and the speed differences are getting higher. Many bicycle facilities are quite narrow. At some point, these facilities need to be of higher quality to handle the speed differences and diversification of cycle traffic. The good thing is that policy makers around the world can learn from countries with higher bicycle shares (like the Netherlands) and be more proactive. 

Interested in learning more? 

You can contact Roxy at roxytacq@icloud.com 

https://www.roxytacq.com/