Archived: London Collision Map and Improved Cycle Journey Planning

As part of a wider road safety strategy, last week we launched the London Collision Map and London collision data on our website and in our API, as part of our ongoing commitment to providing open data. We’ve also made improvements to cycle journey planning.

The new cycle journey planning features include:

  • Google street view images at every turn of the route to help cyclists prepare a ‘mental map’ of their route and visualise junctions before they make their journey.
  • The location and details of cycle parking at London rail stations including information on the number of potential spaces, the type of cycle parking and whether it is covered.
  • ‘Cycle route classifications’ for each section of the journey describing the type of cycling environment that cyclists will encounter along their route (e.g. Cycle Superhighways or routes through parks) helping them to be more informed about the journey they will be making.
  • Identification of steep hills along the route.

The other new feature, the London Collision Map, allows searches for road collisions across London, providing information about when and where they occurred, as well as the severity of the incidents, dating back to 2005. It uses STATS-19 casualty dataset, collected annually by the Department for Transport, and shows a significant reduction in the number of collisions over time.

Through this data we’re able to inform people about junctions and roads that have high numbers of collisions, and where road users should be particularly careful. The 2014 figures show that this number fell to its lowest level since records began, in line with the Mayor’s target to halve the number of people killed and seriously injured (KSI’s) by 2020.

The London Collision Map, showing fatal and serious collisions in 2014, around the area of St. James's Park Underground station.
The London Collision Map, showing fatal and serious collisions in 2014, around the area of St. James’s Park Underground station.

The tool allows for filtering by “Casualty type” (i.e. the parties involved), the severity of the collision, the age of the injured party, and the year that the collision took place. Leaving all the filters at the defaults, and entering “St. James’s Park” in the search box, we see the example above. There are serious incidents on Whitehall and Birdcage Walk, as well as other clusters of incidents, such as the group of seven around Parliament Square.

Details of the collision are available by clicking on the map marker. Notice also that the clusters are broken up as the map is zoomed in. (You can also break up the cluster by clicking on the cluster’s marker).

Selecting a marker shows the details of the collision.
Selecting a marker on the map shows the details of the collision highlighted.

To access the raw data behind the tool, you can use the AccidentStats endpoint, which returns data a year at a time. For example, 2014:
   "location":"St Botolph Street junction with Houndsditch",
   "borough":"City of London",
   "casualties": [{
   "vehicles": [{
// etc.

In the JSON extract above, it can be seen that the data structure allows for multiple vehicles and casualties to be involved in a collision. The possible values of the enumerations in the data are detailed below.

Severity: "Fatal", "Serious", "Slight"
Casualties.AgeBand: "Adult", "Child", "Unknown"
Casualties.Class: "Driver", "Passenger", "Pedestrian"
Casualties.Mode: "BusOrCoach", "Car", "GoodsVehicle", "OtherVehicle", "PedalCycle", "Pedestrian", "PoweredTwoWheeler", "Taxi"
Vehicles.Type: "AgriculturalVehicle", "BusOrCoach", "Car", "HeavyGoodsVehicle", "LightGoodsVehicle", "MediumGoodsVehicle", "Minibus", "Motorcycle_0_50cc", "Motorcycle_125_500cc", "Motorcycle_500cc_Plus", "Motorcycle_50_125cc", "OtherMotorVehicle", "OtherNonMotorVehicle", "PedalCycle", "RiddenHorse", "Taxi", "TramOrLightRail"

We realise that the collision map is just one way of interpreting the data, so we’ve made all of the raw information available in the Unified API as we continue to make our open data provision as comprehensive as possible. We really look forward to seeing the ways in which our developer community are able to use this data, and what kind of tools are created from it. If you’re working with this data, please let us know in the comments below what you’re working on.

Also, with this data set we have taken a slightly different approach to our other APIs, making the data available in large chunks (by year) rather than offering lots of filter options on the endpoint. Do let us know in the comments which approach you prefer for this type of API.

1 Comment

  1. Mr. Williams

    Fascinating article! As a university professor, I plan on using the TFL API in my programming course. Also, I’ll be bringing a group of university students studying I.T. through London this August and would love to visit your facilities so they can witness first hand how the TFL makes this happen. If this is a possibility, please contact me! mafudge at syr dot edu


Comments are closed.