How to use and contribute to Melbourne cycling safety data ~ Better By Bicycle

Summary: Cycling crash and black spot maps and other safety visualisations capture a lot of attention but rarely provide individuals with accurate, self-explanatory means to determine the safest routes or best mitigate actual risks. This post will collate the various data, maps and apps for Melbourne and provide guidance on how individuals can best get value from this data and also contribute to improving cycling safety.

CrowdSpot: Cycle Stories

Related Posts:
> See the safety and health category of this blog
> Where are Melbourne's real cycling danger spots?
> The facts about cyclist deaths on Australia's roads

Details:
1. What cycling safety data is collected for Melbourne? How complete is it?
- I've collected the main, ongoing collections of cycling crash, incident and safety data below. The biggest limitation of cycling crash data collection in Australia is that incidents are captured only if police attend and file a report or a police report is filed later. Thus safety risks and incidents that don't cause serious crashes are typically not captured. Even incidents when an ambulance attends are not captured unless police are involved.

See:
> VicRoads: Crash Stats
> VicRoads: - Crash Stats - Data Extract
> VicRoads: Cyclist safety
> VicRoads: Bike rider safety
> BITRE - Australian Road Deaths Database
> CrowdSpot: Cycle Stories (crowdsourced)

Some of the best uses of the data that is collected are as aggregate statistics of the major causes of cyclist casualties which can provide some insight as to common risks:

Bicycle crashes are most common (61%) at intersections.
12% of bike rider casualties occur when a driver or rider who is turning right fails to give way to an oncoming vehicle travelling straight through.
10% of bike rider casualties occur when a driver or rider fails to give way at a cross section.
10% of bike rider casualties occur when a driver opens their door into the path of a bike rider. This is much higher in inner Melbourne.
9% of fatal and serious injury to bike riders occurs when it is mainly the rider coming off the footpath.
8% of fatal and serious injury to bike riders occurs when it is mainly the car driver emerging from a driveway.
See: VicRoads: Bike rider safety

2. What maps/visualisations, apps and analysis of this cycling safety data currently exist?
- I will collate the existing, accessible maps/visualisations below. However, the main limitation with most maps is that they don't chart the crash rate (accounting for route popularity) thus these maps mostly tell you simply where cyclists currently ride rather than where the elevated risk is. They do confirm that the vast majority of cycling crashes occur at or near intersections though.

(a) Monash University: City Science has created data visualisations based primarily on the VicRoads: Crash Stats from July 2008 to July 2013.

> Melbourne Bike Crash Map (2008-2013) - Monthly crashes
> Melbourne Bike Crash Map (2008-2013) - Map
- Provides a useful map of Melbourne bike crash data (July 2008 to July 2013) with filters for Fatal crashes, Serious injuries and Time Window. It also helpfully integrates Strava data showing where cyclists most commonly ride (skewed somewhat to recreational and experienced cyclists).

(b) Trip Risk
- Provides crash statistics from July 2007 to June 2012 for walking, cycling and driving based on
specific routes between user-selected points. Unfortunately, the route can't be refined but the breakdown of the crash data (Time of day, Day of week, Age, Crash type) is interesting.

- A purely crowdsourced map of cycling crashes and incidents which allows for extremely useful details of how each crash happened.

- There is plenty to learn from many of these crashes. If you ride regularly on the same route (e.g. your work commute) it would be very productive to review the details for all crash spots on your route so that you can learn about the risks and take specific precautions. For example, here is a common black spot on a popular commuting route that can readily be minimised by deliberately slowing down on this downhill stretch and ensuring there are no cars turning or pedestrians crossing:

- I've used CrowdSpot data extensively to provide specific advice on how to mitigate common cycling crash risks in these posts:

> Where are Melbourne's real cycling danger spots?

> Improving Brunswick St and St Georges Rd south of Merri Parade

(d) The Age: Black spot map

> Cyclists get warning on danger spots
> Melbourne's cycling black spots map

- These black spot maps suffer from various deficiencies, particularly: (a) When they don't account for route popularity; (b) Because they reflect past crashes only and not improvements making cycling safer in the high-crash areas and (c) Because some types of crashes are either entirely avoidable or can be significantly reduced (e.g. dooring).

- However, the biggest issue with most maps of crashes or danger spots is that they don't tell people the recommended safe routes that they should ride, which is what they really want to know. For example, simple perusal of the cycling crash maps will indicate that Sydney Rd (especially between Albion St and Brunswick Rd) is a black spot. However, these crash spots don't indicate to novice cyclists that the adjacent Upfield bike path is a very safe and recommended alternative. Or that Royal Parade (the continuation of Sydney Rd) is also safe enough. On the map below you can see the routes with many crash spots versus the green recommended routes (including off-road and bike boulevards) that I have marked up.

See: Melbourne map of key cycling transport routes, infrastructure and destinations

(e) Bicycle Network RiderLog vs Strava

- RiderLog is a smartphone app that logs ride data in a similar way to Strava but uses it for bicycle infrastructure planning. It has only been available for iPhones till now but Android and Windows versions are coming in December 2014. Given the historical lack of making the collected ride data available in useful, transparent ways online (like Strava Labs) I recommend transport cyclists use Strava to log their rides.

3. Useful insights from aggregate cycling crash statistics and visualisations
- Generally I find the simplistic crash maps don't deliver very useful insights for individual cyclists seeking safer routes and to minimise risks from known dangers. However, there are a few exceptions.

- Using the Time Window filter on the Melbourne Bike Crash Map (2008-2013) - Map to choose the peak work commute periods (e.g. 7am - 9:30am) does indicate that there may be some higher risk areas on popular cycling commuter routes to Melbourne CBD - I've highlighted some of these sections below. When combined with actual experience and reviewing specifics from CrowdSpot data you would conclude that there are some relatively higher risk spots on the very popular St Georges Rd route that are worth paying attention to. However, the section of Sydney Rd (and Sydney Rd generally) is worth eliminating from a regular commute if feasible (e.g. using the Upfield bike path), though Royal Parade is quite safe.

- Conversely, much critical data lacks detail and useful context on such maps. For example, filtering the fatalities on such maps provides little insight into areas of high risk as there is insufficient information about the crashes, circumstances and causes of the fatalities.

4. Pursuing serious injuries and fatalities with the police and state/local government is the most effective way of achieving significant change

- In terms of media attention, political responsiveness, systemic government accountability, funding and achieving real change, a single high-profile fatality or serious injury that is effectively pursued is worth 1000 emails and crowdsourced contributions. Even police reports or submissions to government about improving cycling safety typically have limited traction unless linked directly to fatalities or serious injuries.

- The tragic death of James Cross due to being doored on Glenferrie Rd in Hawthorn in 2010 is a perfect example of this. Cyclists and cycling safety advocates campaigned for years before this to reduce dooring risks and fix "door zone bike lanes" with little action from the state or local governments. Dooring was generally treated as a trivial issue by government and police in the larger road safety universe. The death of James Cross is what really changed the response to dooring and is the context for broad action across Melbourne in creating safer new bike lanes or improving existing ones.

- The aggregate Crash Stats safety data has little to tell us about such incidents. Below is the crash data point for James Cross' incident. These data points and maps tell us nothing about the crash specifics or even in aggregate which routes have the highest risk of dooring.

- You can see from the actual standard crash metadata collected why these aggregate crash statistics have such limited value. The crash data recorded for the death of James Cross is below. Reading it alone, one would assume James had a direct accident with a truck rather than being doored and flung into the path of a truck.

- It's hardly encouraging for individuals to spend their precious time/effort filing police reports for each minor dooring (or near dooring) when even if their data point was recorded it is likely not going to be used to achieve safer cycling. The sheer weight of dooring reports isn't going to drive rapid improvement even after James Cross' death. After all, VicRoads now provides alarming statistics about dooring but still puts the onus on drivers and cyclists rather than safe cycling infrastructure:

One of the biggest risks to bike riders is car doors being opened into their path. The bike rider may swerve out further into the road or collide with the car door, often with serious consequences. Between 2007 and 2011 there was an average of 38 serious injuries per year as a result of car door and bike rider collisions. There has been one fatality in the last five years. This is eight per cent of all bike rider serious injuries and fatalities.
Advice for riders
Look out for drivers and passengers getting in and out of parked cars.
Be vigilant when riding alongside parked cars and ride out of the car door zone (if possible and safe to do so).
If you are riding on a length of road with a marked bicycle lane, you must ride in the bicycle lane unless it is impracticable to do so.
In places where there are a lot of parked cars, slow down.
Wear bright coloured clothing and use lights at night or in conditions of low light.
VicRoads: Car doors & bike riders

- It's clear that the most efficacious approach is to persistently follow up serious injuries and fatalities until systemic improvement is achieved. Those serious injuries don't have to be your own, they could be friends or just residents of your community. All cyclists faced with dangerous "door zone bike lanes" should be holding VicRoads, local and state governments and politicians accountable and mention James Cross and any serious dooring injuries of people they know. The same applies to all of the other common types of serious cycling injuries.

James Cross was the 13th cyclist to be killed on Victorian roads in the past two years but the first recorded death by ''dooring''. The inquest on his death in March last year heard that the driver, Ellen Richards, 60, had looked in the driver's side mirror before opening the door about 12 centimetres. Mrs Richards said she did not see the cyclist who struck the door as it opened. Coroner Heather Spooner said Cross' death was entirely preventable. ''It has highlighted a very significant public safety hazard, particularly in high-risk areas where car dooring is responsible for many injuries to cyclists,'' she said. Cross' death has now become a rallying point, provoking action. Reforms are being planned to make Melbourne safer for cyclists and to change motorists' behaviour towards riders. Moves include:

Bicycle Network Victoria has asked the police Chief Commissioner to ensure drivers involved in dooring incidents are charged and fined.

VicRoads will launch a campaign in early 2012 aimed at encouraging ''understanding and respect between cyclists and drivers to share the road safely and mindfully'' with a focus on dooring.

RoadSafe Action Group (a partnership of four inner-city councils) will begin Operation Doorknock in 2012 to encourage drivers to adopt new safety techniques when getting out of parked cars.

The Amy Gillett Foundation is lobbying state and federal governments to include bike-related safety issues in driver training and licence testing procedures.

James Cross' parents, both doctors, hope that lessons can be learnt from their son's death. ''The message we want to give is one of driver awareness and mutual respect,'' says his mother. ''Drivers must be absolutely aware and careful every time they get into and out of a car, and councils and government have a responsibility to make roads and bike lanes safer through road infrastructure and driver education. These problems aren't going to go away. They need to be addressed before another life is lost,'' she says.
Cyclist's death a catalyst for change

- Nevertheless, the types of improved cycling infrastructure and safety improvements have to be evidence-based and actually improve safety and convenience for cyclists. Spending a little money in just the area of the fatality or serious injury to do something different (not necessarily significantly better) is not the desired outcome. If you depend on cycling for transport you need to identify the best, feasible solutions (not settle for bandaid/token measures) and persistently advocate for them - particularly with your local and state government politicians.

Garry Brennan, of Bicycle Network Victoria, cautioned that the lane was ''experimental'' and could create confusion. ''We are concerned that drivers could assume that bikes are only permitted to ride in the coloured half of the bike lane, and drivers might mistakenly think they are free to open doors across the section of the lane that is not coloured,'' he said. ''Glenferrie Road certainly needs better bike infrastructure, but at this stage there is no evidence to suggest that this proposed design will be effective.''
Trial hopes to shut the door on bike accidents