It seems pretty much anywhere cyclists like to ride, either locals or non-cyclists who like to drive there are willing to fire up a petition. The latest picked up traction on the bike-hating (and oddly still title sponsor of the Sun Tour) Herald Sun.

Since I haven’t for a while, I figured I’d take a look at the actual data and give it some evidence-based analysis:

I’ve covered most of the process in a previous post, so won’t go over the details again. However, a few notes are worth mentioning since the dataset has been updated since I downloaded it last:
  1. There are now lat-long coordinates in NODE.csv so no need for DIY projections, (which is nice)
  2. There’s something odd with with VEHICLE_TYPE in VEHICLE.csv which indicates there were only 4 accidents involving bicycle1.
  3. As a result I used Road User Description in PERSON.csv to filter cycle v non-cycle accidents.

How does the data plot out? The following image covers all accidents in the 2006-2016 data extract for the area of the Shire of Yarra Ranges. It looks pretty nasty with 4,312 accidents in the 10 year period ~ an average of just over 1 per day.

Accidents - all

Out of these 4,312 points, how many are created by cyclists? Surely with proposals they are ‘banned for their own safety’ it’s going to be pretty impressive. Afterall; tons of cyclists head that way for the scenery and the climbs.

The Strava heatmap below shows some of the popular areas around Melbourne with the Yarra Ranges showing up nicely north east of Yarra Junction.

Strava Heatmap

It’s not quite beach-road busy, but still pretty popular. Annoyingly the Strava Metro data sample doesn’t reach this far east so I can’t good numbers, but a quick check on segments in the area gives 700-800 efforts this year so far – extrapolate that as you will.

A quick definition query on the full dataset yielded the terrifying extent of cycling accidents in the area.

Accidents - cyclists

All 114 of them. In 10 years. For those who are mathematically challenged, an average rate of 11 per year (or one accident per month).

The non-cyclists take care of the other 4,198 incidents, averaging 4 fatal crashes and 88 serious accidents every year. It’s not even worth calculating the ratio of cyclist to non-cyclists.

Accidents - noncyclists

From the comical petition which has so far garnered 97 signatures:

Petitioner Larry Colwell said in the change.org petition that it was not “anti-cycling” and it aimed to “prevent death and injury for all road users”.

Injury Severity All Non-cyclists Cyclists
Fatal 40 40
Serious 1,108 1,062 46
Minor 1,812 1,747 65
None 1,352 1,349 3
Total 4,312 4,198 114

I think I have reasonably demonstrated that the best way to prevent injury and death would be the banning of motor vehicles on these roads, which is as stupid as it sounds. At best, the proposed cycling ban will cut the average monthly accident rate from 36 to 35.

I didn’t feel like doing a year-by-year analysis, but all the tools and data are free/ Open Source, should you wish to.

  1. That’s right, unlike some people I actually check the data against what I seem reasonable, instead of just blindly plowing through.

Posted in General,GIS | Tags: , ,  | 1 Comment