We have access to more transport data than ever before. We collect crash data, traffic volumes, speed limits, road characteristics and vehicle speeds. Increasingly, we also have access to large, connected vehicle datasets showing how vehicles actually move across entire road networks. The coverage and quality of this data varies, but the opportunities to make better use of it are growing.
I spent a lot of time thinking about this while working on the recently updated Austroads Guide to Road Safety Part 3: Safe Speed Management (AGRS Part 3), particularly the new chapter on data, monitoring and evaluation we developed. One of the things that struck me is just how quickly this area is changing. We have much more data available than ever before, but we’re still learning how to make best use of it.
Traditionally, understanding vehicle speeds meant putting out tube counters at individual sites. This method remains valuable, particularly where we need accurate information about what is happening at a specific site. However, it only gives us part of the picture.
Probe data derived from vehicle telematics, GPS devices and other sources can provide speed information across much larger parts of the road network. Importantly, it can also give us information retrospectively. If a speed limit changed six months ago and nobody commissioned a ‘before’ tube count, there may still be data available to help understand what changed. Probe data can also tell us how speeds changed along a corridor, rather than just at individual sites. This is valuable where speed management infrastructure treatments are implemented across a wider network, for example through Local Area Traffic Management (LATM).
For me, the network view is where things start to get interesting. We can increasingly move between looking at an individual road and looking across an entire network. That opens up possibilities for identifying where actual speeds are inconsistent with the desired ‘safe’ speed, prioritising interventions, and understanding how speeds vary by road environment.
At Abley, much of our work already involves bringing different spatial datasets together to understand transport networks and risk. Speed data adds another increasingly powerful layer to that picture.
More data doesn't automatically translate to better decisions.
Every data source measures something slightly different, so getting value from it depends on understanding what it tells us. Probe data can provide speed information across whole networks and over time, rather than just at individual survey sites. But it represents a sample of vehicles, coverage may be limited on low volume roads and speeds are measured differently from tube or radar surveys. This does not make one approach better than the other. Each has different strengths and limitations, and the most useful source will depend on the question being asked.
Making good use of the data also means starting with the question rather than the dataset. Are we trying to:
Identify parts of a network where vehicle speeds are consistently high?
Understand whether a speed limit change actually affected travel speeds?
Identify where engineering treatments might be needed to support a speed limit reduction?
Estimate the likely safety effect of a change in mean speed?
Understand whether concerns about increased journey times following a speed limit change have materialised?
These are different questions, and they do not necessarily need the same data or the same level of precision.
The updated AGRS Part 3 reflects this broader range of applications. Tools such as the NZ Transport Agency's MegaMaps spatial application and the Global Road Safety Facility's Speed Impact Tool demonstrate different ways speed-related data can be turned into something practitioners can actually use.
Let’s be honest, it always is though, isn’t it?
This is probably my biggest takeaway from working on the data section of the AGRS Part 3. Too often, monitoring gets considered once something has already been built or changed. At that point somebody inevitably reflects: ‘I wonder if that actually worked?’ Ideally, we should have decided how we were going to answer that question before implementation.
For speed management, the immediate answer often comes from vehicle speeds. Crash outcomes obviously matter, but statistically reliable before-and-after crash evaluations generally require several years of data.
Changes in vehicle speed can be observed almost immediately, and that gives us a much faster feedback loop. If speeds have reduced to the level intended, great. If they haven't, that tells us something too. Perhaps the speed limit needs greater reinforcement through road design, traffic calming or enforcement.
Sometimes what we want to understand isn't safety alone. We might also want to know what happened to journey times, traffic patterns, walking and cycling, community perceptions or other local outcomes. The AGRS Part 3 deliberately broadens the concept of monitoring for speed management beyond speeds and crashes for that reason.
Speed management is often controversial. You can have excellent data, a sound methodology and a beautifully constructed map showing exactly why a change makes sense, and still not convince everyone. One of the lessons from the broader work behind the Guide is that facts need context.
A graph showing a reduction in mean speed is useful. Showing what that reduction means for the likelihood or severity of a crash makes it more meaningful. And sometimes a local story about a child walking to school, somebody trying to cross a busy road or a community wanting a safer main street makes the issue understandable in a way that another graph simply won't. The AGRS Part 3 similarly recommends combining evidence with values-based and local storytelling when communicating about safe speeds.
Good data gives us confidence that we understand the problem, but a story helps us explain why solving it matters.
What excites me most about this area is that I don't think we've figured out the possibilities yet. Probe datasets are improving and spatial datasets are becoming richer. Web maps and dashboards sharing speed-related data and insights are making complex analysis more accessible. We can increasingly analyse speeds across whole networks rather than isolated survey sites, and we can link that information with road risk, infrastructure, land use and other data.
The AGRS Part 3 acknowledges that the quality and resolution of probe data are continuing to improve, and that some of today's limitations may not remain limitations in the future. It also encourages evaluations of emerging interventions and data sources to be shared, rather than remaining within individual organisations.
There is a lot more we can do with the information already at our fingertips, and I suspect we are only just getting started.