Pedestrians hit by vehicles while on the pavement - the data

Yesterday I posted about pedestrians hit by vehicles on pedestrian crossings and how the vast majority of those are motor vehicles, despite a perception that this is a particular problem with cyclists. Today I am looking at pedestrians who have been injured while on the pavement, also a common complaint about cyclists.

The map below shows the vehicles which have hit pedestrians on the footway, verge or refuge/central island in the years 2005-2015.

In this case a shocking 97% of pedestrians hit on the pavement were hit by motor vehicles, with just 3% of the injuries attributed to cyclists. This presumably includes 'shared space' pavements where cycling is permitted and encouraged.

Again this highlights that if local councillors are serious about pedestrian safety they need to focus on the danger caused by motor traffic. I would also encourage them to reject the many poorly designed and poorly signposted shared space pavement schemes, which even though it does not result in many accidents, often gives the impression cyclists are being inconsiderate when they have been encouraged to use the pavement.  

Key:

Car - yellow marker (there isn't a car symbol)

Bus, coach or minibus - bus icon

All vans and goods vehicles - truck icon

Motorcycle - motorcycle icon

Pedal cycle - bicycle icon

Taxi/Private hire car - cab icon

For more detail on any of the data points simply click on the icon.

These maps are based on the data STATS19 police collision report data from 2005-2015 inclusive (11 years of data). If you want the full details on specific a specific incident or want to search by area I recommend using www.cyclestreets.net/collisions/.

Pedestrians hit on crossings - the data

This map shows the vehicles which have hit pedestrians on pedestrian crossings 2005-2015.

Last week one group identified cyclists as a main cause of road danger. A brief look at road casualty statistics shows this is not the case.

For instance, cyclists jumping red lights are often cited as a problem on our roads. However if we focus in on this 'cycling' problem, the road accident data shows motor vehicles are responsible for the vast majority of collisions with pedestrians on crossings. 

Only 2% of the pedestrians hit on crossings are hit by cyclists. That is two percent too many, but if we are serious about pedestrian safety we need to focus on the 98% of collisions which cause the most serious injuries and deaths.

If councillors are serious about avoiding conflict between cyclists and pedestrians they need to stop the widespread use of poorly designed and poorly signposted shared space pavements. Instead they should provide dedicated cycling facilities protected from traffic as far as possible and leave pavements clear for pedestrians. 

Key:

Car - yellow marker (there isn't a car symbol)

Bus, coach or minibus - bus icon

All vans and goods vehicles - truck icon

Motorcycle - motorcycle icon

Pedal cycle - bicycle icon

Taxi/Private hire car - cab icon

For more detail on any of the data points simply click on the icon.

In some locations you will see two icons exactly on top of each other. This is as indicated in the original data and may have been due to locations being rounded or approximated previously. For instance at long. -0.133876, lat. 51.519942 two icons are shown over each other representing separate incidents on 21/02/2008 and 08/07/2009. 

These maps are based on the data STATS19 police collision report data from 2005-2015 inclusive (11 years of data). If you want the full details on specific a specific incident or want to search by area I recommend using www.cyclestreets.net/collisions/.

The danger of 'Agricultural Vehicles' in Central London?

The map below shows where pedestrians and cyclists have been killed or injured in collisions involving 'agricultural vehicles' as recorded on police road collision reports for the years 2005-2015.

The cluster of casualties in central London highlights problems with the current reporting of collisions - vehicles are put in to categories which do not adequately describe their purpose. In this instance it's likely most or all of these 'agricultural vehicles' were in fact diggers and other construction vehicles lumped in to the same category (details here).

This is in addition to the 'other' vehicle category which encompasses many of the most dangerous construction vehicles and a plethora of other vehicles, masking the large number of deaths caused by a these vehicles. We also know from insurers that the construction industry has a poor record when it comes to vans and small lorries.

It's been clear for a long time that the construction industry has a problem when it comes to road safety, unfortunately the data collected hides the extent of the problem. In order to understand the real causes of road danger we need better data to be made available, including detailed vehicle types and maybe also the industry sector.

Pedestrians and cyclists killed or injured in collisions involving agricultural vehicles (2005-2015): fatal - black, serious - yellow, slight - blue. 

Notes:

• If you want the full details of specific a specific incident, or to search by area, try www.cyclestreets.net/collisions/.
• Where possible casualty records have been linked to the vehicle which hit them. For pedestrians, the data records which vehicle (first) hit them. For other road users we inferred a collision where there was just one other vehicle involved. When there were more vehicles it was not possible to infer which vehicles collided with which.
• Care has been taken to avoid errors or emissions while creating these maps, however if you spot an issue please let us know.

Horse Riders killed or injured 2005-2015 | Map

For a larger version of the map, click here.

STATS19 mapping implementation notes

These maps are based on the data from 2005-2015 inclusive (11 years of data), which has been edited down to make the file sizes more manageable. If you want the full details on specific a specific incident or want to search by area I recommend using www.cyclestreets.net/collisions/.

The data for "Journey_Purpose_of_Driver", "Vehicle_Type", "Make" and "Model" refers to the other vehicle involved where that has been possible to match*. For pedestrians, the data records which vehicle (first) hit them. For other road users we inferred a collision where there was just one other vehicle involved. When there were more vehicles it was not possible to infer which vehicles collided with which.

*Bicycles, horses and mobility scooters are all counted as vehicles.

For a few of the entries we have found errors in the data, such as missing location data which may means they have not been mapped. Examples of this are accident references 2005600GN6024 and 200504BB05106.

Care has been taken to avoid errors or emissions while creating these maps, however if you spot an issue please let us know.

Alternative map of KSIs involving Heavy Goods Vehicles

In my previous post I presented a map created using Google Fusion Tables. This post shows an alternative map produced using the same data fed in to Google Maps. Due to limits on the file size, the collisions resulting in minor injuries have been omitted.

Vulnerable road users killed or seriously injured in collisions involving HGVS for the years 2005-2015

For a larger version of the map, click here.

'Vulnerable Road Users' refers to pedestrians, cyclists, motorcyclists, horse riders & mobility scooter riders. Due to the number of dangerous goods vehicles classed as 'Other', that category has been included and the maps will also include some collisions involving different vehicles in the 'Other' category.

STATS19 mapping implementation notes

These maps are based on the data from 2005-2015 inclusive (11 years of data), which has been edited down to make the file sizes more manageable. If you want the full details on specific a specific incident or want to search by area I recommend using www.cyclestreets.net/collisions/.

The data for "Journey_Purpose_of_Driver", "Vehicle_Type", "Make" and "Model" refers to the other vehicle involved where that has been possible to match*. For pedestrians, the data records which vehicle (first) hit them. For other road users we inferred a collision where there was just one other vehicle involved. When there were more vehicles it was not possible to infer which vehicles collided with which.

*Bicycles, horses and mobility scooters are all counted as vehicles.

For a few of the entries we have found errors in the data, such as missing location data which may means they have not been mapped. Examples of this are accident references 2005600GN6024 and 200504BB05106.

Care has been taken to avoid errors or emissions while creating these maps, however if you spot an issue please let us know.

Mapping KSIs involving Lorries

Lorries and other large vehicles are a serious danger to vulnerable road users.

Using the STATS19 data I've produced maps showing collisions involving these vehicles where pedestrians, cyclists, motorcyclists, horse riders or mobility scooter riders were killed or injured. Due to the number of dangerous goods vehicles classed as 'Other' (details here) the maps will also include some collisions involving different vehicles in the 'Other' category.

Deaths & Injuries involving HGVs, 2005-2015

Key: 
Fatality = Symbol
Serious Injuries = Yellow dot
Minor Injuries = Blue dot

This is a map I produced showing where vulnerable road users have been killed or injured 2005-2015. Click here for full size version, produced using Google Fusion tables.

Other mapping:

In addition to this, I have tried presenting the data using Google 'Maps' rather than 'Fusion Tables'. This has several advantages however it limits each layer to 2000 data points, which poses challenges when dealing with data for the whole of the UK (except Norther Ireland). Due to this limitation it is necessary to filter the data down, reducing the years, casualty types or casualty severities shown.

Click here for a map showing where pedestrians, cyclists and motorcyclists have been killed or injured in collisions involving Heavy Goods Vehicles for the years 2012-2015. This is a subset of the data shown in the map above.

STATS19 mapping implementation notes

These maps are based on the data from 2005-2015 inclusive (11 years of data), which has been edited down to make the file sizes more manageable. If you want the full details on specific a specific incident or want to search by area I recommend using www.cyclestreets.net/collisions/.

The data for "Journey_Purpose_of_Driver", "Vehicle_Type", "Make" and "Model" refers to the other vehicle involved where that has been possible to match*. For pedestrians, the data records which vehicle (first) hit them. For other road users we inferred a collision where there was just one other vehicle involved. When there were more vehicles it was not possible to infer which vehicles collided with which.

*Bicycles, horses and mobility scooters are all counted as vehicles. 

For a few of the entries we have found errors in the data, such as missing location data which may means they have not been mapped. Examples of this are accident references 2005600GN6024 and 200504BB05106.

Care has been taken to avoid errors or emissions while creating these maps, however if you spot an issue please let us know.

Croud-sourcing information for collision maps

The police STATS19 reports contain few details about victims and the circumstances of collisions. Linking these to press reports, blog posts & local campaigns can give more information on the people involved, the incident, the investigations and any court cases.

Finding, linking and mapping press reports about all road fatalities in the UK would be a huge task to undertake. Therefore I wanted to look at ways of allowing members of the public to submit reports online in a way which could be easily mapped and linked to other data.

I've created the form below as a prototype - please try it out and let me know what you think.

Mapping Road Danger - Submit Report/Article

The main challenge was getting the location data. Ideally the user would be able to click on a map and have the location recorded. I wasn't able to do that, so the user has to copy the latitude and longitude in to the form - OK on desktop, but really not phone/tablet friendly.

Output

The form populates this table (email addresses removed):

This probably needs a disclaimer 'This data includes links and information submitted by the community, please contact us if you believe any of the information is inaccurate or inappropriate'.

Using the data

• This data could be mapped to provide real time logging of incidents.
• Once the STATS19 collision reports are published it would be possible to link the data sets. These could be matched with a good degree of accuracy using the date and a search radius, however this might still need a final check by a human. 
• Similarly, this could be used to link historic reports to the STATS19 data to improve our understanding of the accident history of locations.
• This form (or similar) could be used by people currently recording collision reports to put the data in a consistent format which would allow it to more easily mapped in future.

Mapping of collision reports is already being undertaken by a few people (here for instance) and if efforts were to be put in to this it would be good to coordinate to avoid duplicating efforts and ensure data could be transferred across if needed.

STATS19 collision data - Vehicle and casualty categories

When the police record road traffic incidents they use various classifications to record details vehicles, casualties and injuries. This information is entered in to the STATS19 database which is published annually.

The STATS19 data can be found here. The details of how vehicles, casualties, injuries etc. should be classified is published on the STATS20 form.

Vehicle Classification

The vehicle classifications are shown below. The number on the left is the code stored in the police database for each category (for the detailed notes about classification refer to the STATS20 link above).

It should be noted that many vehicle types do not fit the classification well - construction industry diggers would be classified as agricultural vehicles and 'other' category contains refuse vehicles and a variety of construction lorries and equipment. As a result, the goods vehicles data significantly under-represents the lorries most often causing fatalities by as much as 10-15% for both refuse lorries and construction lorries (based on TfL data for London 2010-2014). Taken together this could mean around a fifth of lorry fatalities are recorded in the 'other' category.

• 01  Pedal cycle - includes tricycles etc.
• Motorcycles - 'motorcycle' includes mopeds, motor scooters and three wheeled motorcycles.
• 02  Motorcycle 50cc and under
• 03  Motorcycle over 50cc and up to 125cc
• 04  Motorcycle over 125cc and up to 500cc
• 05  Motorcycle over 500cc
• 97  Motorcycle – unknown cc
• 23  Electric motorcycle - includes all electrically powered motorcycles over 0.25 kW, whether or not they are also equipped with pedals.
• 08  Taxi/Private hire car - any vehicle operating as a hackney carriage, regardless of construction.
• 09  Car - includes three wheeled cars, estate cars, family vans/multi-people carriers, Land Rovers and similar four-wheel drive vehicles.
• 10  Minibus (8 - 16 passenger seats)
• 11  Bus or coach - defined as 17 or more passenger seats.
• 16  Ridden horse
• 17  Agricultural vehicle - includes construction vehicles such as mobile excavators and front dumpers, but not heavy plant.
• 18  Tram/Light rail
• Goods vehicles include tankers, HGV tractors travelling without their semi-trailers and articulated vehicles.
• 19  Van/Goods vehicle 3.5 tonnes maximum gross weight (mgw) and under
• 20  Goods vehicle over 3.5 tonnes and under 7.5 tonnes mgw - Sometimes referred to as Medium Goods Vehicles or Large Goods Vehicles. These vehicles are included in the CLOCS safety scheme.
• 21  Goods vehicle 7.5 tonnes mgw and over - Heavy Goods Vehicles (HGVs), although sometimes this term is used to mean all vehicles over 3.5 tonnes.  These vehicles are included in the CLOCS safety scheme.
• 98  Goods vehicle – unknown weight
• 22 Mobility scooter
• 90 Other vehicle -  Predominately the vehicles causing serious injuries and fatalities in this category are large vans, refuse vehicles, tipper lorries, cement lorries and skip lorries. However the category also includes ambulances, fire engines, motor caravans, and quad bikes, plus anything else which doesn't fit the other classifications.

In addition to the information released publicly, the police database includes information such as the registration number. This can be used by the DfT/TfL to determine the exact make & model of the vehicles in order to analyse the data in more detail (especially the 'other' categories).

'Casualty Type' Classification

The casualty classifications are shown below. The number on the left is the code stored in the police database for each category:

0 Pedestrian

1 Cyclist

2 Motorcycle 50cc and under rider or passenger

3 Motorcycle 125cc and under rider or passenger

4 Motorcycle over 125cc and up to 500cc rider or passenger

5 Motorcycle over 500cc rider or passenger

8 Taxi/Private hire car occupant

9 Car occupant

10 Minibus (8 - 16 passenger seats) occupant

11 Bus or coach occupant (17 or more pass seats)

16 Horse rider

17 Agricultural vehicle occupant

18 Tram occupant

19 Van / Goods vehicle (3.5 tonnes mgw or under) occupant

20 Goods vehicle (over 3.5t. and under 7.5t.) occupant

21 Goods vehicle (7.5 tonnes mgw and over) occupant

22 Mobility scooter rider

23 Electric motorcycle rider or passenger

90 Other vehicle occupant

97 Motorcycle - unknown cc rider or passenger

98 Goods vehicle (unknown weight) occupant

Mapping for road safety

For road safety campaigners much of the information we have could be mapped, whether that's speed limits, consultations/schemes or collisions. With such a lot of information, no one maps will show everything. Maps for communicating with the public tend to be the simple, while those to aid decision making often have multiple layers of information.

I've been reviewing various sites to identify what information is already being mapped to identify how it might be used for campaigning (or whether there are maps we're missing).

Police Collision Reports

There are several websites which map the UK wide STATS19 database of police 'accident' reports in different ways. The data is published annually for the proceeding year at in late June (details here):

• crashmap.co.uk A mapping tool based on supported by the ability to buy detailed reports. Can filter by severity, casualty type and year and includes information explaining the definitions used.
• map.itoworld.com/see-me-save-me [HGV, cyclist & pedestrian, KSIs] displays the STATS19 data from 2000-2010 with the casualty's age, gender and the year of the accident shown graphically. Discontinued
• collisionmap.uk a basic map which displays limited information and can only filter by casualty severity.
• cyclestreets (see below) maps all road users and has the full information.

Going beyond Police reports

The following websites map other data related to road safety:

• collideoscope.org.uk/ [cycling, incidents, UK] Collisions and near misses, plus the ability for users to add their own experiences. Also includes STATS19 record information. MySociety/Barts Charitys Safe Cycling Appeal initiative.
• Fill That Hole [Cycling UK / CTC, road defects, UK] A map showing reported road defects with indicators for whether the issue has been fixed or not.
• cycling-intelligence.com [cycling, KSIs, London]  Cyclist fatalities in London updated regularly and includes some information and links. The map goes back to 2006, although they say the data is not complete for the early years.
• beyondthekerb.org.uk/casebook [cyclists, KSIs, UK, 2015-2016] Details, map and links to press reports.
• cycleinjury.co.uk/map STATS19 map combined with information from Levenes personal injury lawyers.
• Cyclestreets [cycling, UK]  In addition to mapping the STATS19 information with detailed search/filtering it has a mapping tool for campaigns, consultations and issues used by cycling groups around the UK. Allows for issues to be identified and discussed. Issues in London can be seen on the LCC website.
• Give a beep [cycling, London] is a scheme where cyclists are given a button on the handlebars which registers the time and location every time it is pressed. The intention is that users press it every time they feel less safe and this builds up a map of the problem locations around the city for use as a campaigning tool.
• Camden Cyclists have produced a similar list of maps including their own analysis of collisions in Camden, plus maps for traffic flows, cycling and mapping specific to London.
• alexinthecities.co.uk - Alex has produced several really useful maps including a map of DfT traffic counts, a list of sites mapping census data and a lot of more London specific maps. He also has information about producing maps using Google Fusion tables. 
• Please let me know if there are any I have missed

Lists of information

There are also several websites listing incidents and details:

• twowheelsandbeyond.blogspot.co.uk [cycling, KSIs, UK]  Cyclist obituaries listed from around 2000-2012.
• visionzerolondon.wordpress.com [cyclists and pedestrians, KSIs, London] Names where known, links and some photos 2014+
• lcc.org.uk/pages/news is the blog for London Cycling Campaign. Articles about specific locations are tagged and can be viewed on a map. An RSS feed of the data is available.
• The campaign group See Me Save Me are collecting reports of collisions and fatalities involving HGVs and are planning to publish this in the future.
• Again, please let me know if there are any I should add.

Conclusions

There is a lot of good mapping being done on the STATS19 data, but the story behind the data (news reports, people's experiences, etc.) are not generally being mapped. There are several initiatives mapping cycling in London but this is quite fragmented.

Creating maps with more detailed information and analysis could be useful to make the case for action to be taken at a particular junction, or to brief journalists following incidents, so would be a useful tool.

For cycling, Cyclestreets seems to be the main focus, however the aim would be to have a UK wide system which covered all incidents with different organisations able to filter for the content relevant to their campaigning.