For this blog entry, we welcome Richard Owen, CEO of Agilysis, a specialist road safety consultancy based in Oxfordshire.
Richard contacted us and very kindly agreed to run his own report on traffic speeds along the A113 through Little End, Stanford Rivers. What followed did not surprise us but it did shock us. If you’re a user of this road, this should shock you too, even more so if you’re a resident.
You can read Richard’s findings below and also learn a little on how the data is gathered and used.
About Agilysis
We have been accessing and analysing connected vehicle data for over five years now, creating our ‘Speed Compliance Tool’ for Slough Borough Council in 2019. There are many different suppliers of data relating to vehicle speeds and they in turn source data from many different places.
The best quality data is derived from connected vehicles, using in-vehicle GPS sensors such as sat-navs, or perhaps tracking devices. Some suppliers with use lower-quality sensor data such as mobile phones although these are typically fine if data is granular enough. Ideally we would like high-frequency data, a GPS ‘ping’ every second would be great but as long as it’s only 2-5 seconds then that’s acceptable. Longer frequencies reduce the ability to get accurate speeds around road features such as bends or junctions, and certainly not at the end of speed limits.
Data from insurance ‘telematics’ devices (black boxes) is less than ideal for monitoring real life speeds as they are likely to be suppressed due to driver behaviour. The argument goes that if you are being monitored, tracked and penalised financially then you’ll likely be amending your driver behaviour so it’s not a true reflection of how most motorists would usually drive.
The Speed Compliance Tool
Our Speed Compliance Tool is now used in over 40 different parts of the country and contains pre-analysed data for the most recent year for each road section. The road network data are provided by Ordnance Survey (OS) with the raw GPS data matched to individual road sections once acquired from the original supplier.
This means you can’t look at speeds for an individual point on a road, but you can look at it for an individual section. Road sections are typically 100m in length but can be longer on rural roads where there are long gaps between significant road features such as junctions and property entrances. For built-up areas the network is typically quite complex and sufficient to give a detailed speed profile along a road or thorugh a village.
Our customers use the data for a whole variety of reasons, mainly to assist with answering queries from members of the public who are concerned about vehicle speeds. It means that authorities don’t need to commission expensive surveys and can simply look up the most recent data online. It also gives them a breakdown of speeds by time of day for busier roads. On many residential streets there doesn’t tend to be enough traffic to get a large enough sample size for analysis though.
Findings
When I heard about the A113SpeederBot project I was impressed and immediately wanted to check to see what data we had for the same stretch of road though Little End. The map to the left shows the section of the A113 through Little End that is subject to 30mph. The ends may not be 100% on the terminal signs as they are matched to the aforementioned road sections provided by OS. The colour coding reflects the difference between average speed (mean) and the posted speed limit. It’s pretty straight forward, green means the speeds are at or below the limit, orange and then red means they are above (colour threshold at 5mph above the limit when it goes to red).
I had a look at the village on Google Street View and I’m not entirely surprised by the results. Vehicle speeds are typically higher where there are only dwellings on one side of the carriageway as driver’s don’t perceive the need to reduce speeds. Ideally speed limits should be ‘self-explaining’; i.e. laid out in a way that encourages drivers to regulate their speeds naturally, and where this isn’t the case other measures (enforcement) need to be put in place to reduce speeds. Without some form of road engineering or enforcement, vehicle speeds will never be close to the posted speed limit at either end of the village.
The series of gauges below is also taken from our Speed Compliance Tool. Is shows the typical speeds at different times of day across the 11 individual road sections shown in the map above. The average speed is 33.6mph and 85th percentile speed a whopping 41.5mph. This means that 15% of drivers are travelling at 11.5mph above the limit along the entire length of the village! We have already seen that speeds reduce towards the centre which means speeds at either end will be much, much higher. Typically police forces will consider requests for enforcement if the 85th percentile speed is above a certain threshold, normally 35mph in a 30 limit. These practices differ around the country and I have no insight into how it works specifically in Essex.
and in table form:
| Time period | mph | 85th |
|---|---|---|
| All Day: Midnight to midnight | 33.6 | 41.5 |
| AM Peak: Weekdays, 0700 to 0900 | 33.2 | 40.4 |
| Off Peak: Weekdays, 1000 to 1600 | 32.7 | 41.2 |
| PM Peak: Weekdays, 1600 to 1900 | 34.5 | 41.9 |
| Evening: Everyday, 1900 to 2300 | 36.5 | 44.1 |
| Night-time: Everyday, 0000 to 0400 | 37.8 | 44.8 |
| Weekend: Weekends, 0700 to 1900 | 34.4 | 42.7 |
Unsurprisingly, we see much higher speeds at evenings and weekends. The off-peak has the lowest average speed but this is still significantly above the limit. There are other sources of data out there that allow even more insight into driver behaviour.
One other finding that became clear was that it can be estimated that Little End sees 10,000 speeding vehicles a day; that’s around 2,000 mid village and 8,000 in the more rural sections, although all of these vehicles are in the 30 mph zone. These are modelled estimates rather than observed figures but given the traffic flow on the A113 they won’t be out by too much.
We have used data from TomTom in the past to create very detailed profiles of speeds on an hour-by-hour basis or even for a specific day, perhaps related to a sporting event. As access to this type of data grows we are expecting authorities to be much more well informed about speeds on their road networks, and having access to tools like this means they can identify issues quickly and prioritise resources.
We’d like to thank Richard and Agilysis for taking the time to lend their expertise in road safety and produce the above post.
Want a heads up on fresh stuff like this? Pop your email address in the box below. Unsubscribing is easy.