On your bike... to where?

The great bike rush of 2020 is in full swing! COVID-19 restrictions means bike shops are sold out and there's a huge increase in cycling everywhere... or so media (social and otherwise) would have you believe, but I haven't seen any details on what is actually happening. So as a cyclist and data scientist I thought I would try and find out.

I wrote a deconstruction of the reported 700% increase in cycling in Cycling, the new COVID-19 side effect? where found some serious sampling bias issues and proposed a more realistic results of 10-150% increase on weekends and a reduction on weekday at most location. While far more robust (in my option anyway) it still did not provide extensive coverage and did not look at changes in where people were riding.

So here we are for round 2 so lets try and make it slightly scientific. My hypothesis (theory of what's happening) is that over the lockdown period:

1. weekday cycle counts will be lower and weekend counts remain unchanged or higher, and
2. weekday peak commuter periods (proportion of cyclists) will be less pronounced or not exist.

The second hypothesis is used as an indication that there might be a reduction in commuter cycling even larger than the (hypothesized) reduction in weekday cycle counts.

If I were to sketch these it would look something like this (1. on left 2. on right, and yes I have the worst handwriting in the world):

Once again I'm using data from the department of transport cycle counter sites but this time I'm using 9 sites (in red) to provide a little more coverage while still being able to look at sites individually.

The sites are mostly off road bike paths however there are a few on-road bike lane locations including Brighton Road, Napier Street and Moreland Street where the counter is located just after the path transitions to an on-road bike lane.

The plot of daily cycle count over time shows some really interesting patterns. Firstly the Christmas period drop is very evident at almost all sites, only Koonung Trail and Main Yarra Trail are exceptions, possibly due to a lower proportion of commuter cyclists, note that they also have similar level of weekday and weekend traffic. After that it is interesting to note that almost every line shows a major change around the time of stage 2 lockdown for Victoria on the 25 March 2020.

While the Christmas dip can make it hard to interpret I think that 6 of the sites displayed a decrease in weekday cycling around the time of stage 2 lockdown with many of them also showing an increase in weekend cycling. The notable exceptions are Gardiners Creek, Koonung Trail and Main Yarra Trail which had a spike in cyclist numbers before a decline in May and June but all of which saw an increase in weekend cycling.

One other observation I had was that the weekend cycling seems to drop off in early to mid June. While we don't have enough data to know if this is a sustained trend but I wonder if recreational riders are getting a little too cold...

So we saw a reduction in weekday cycling at most sites, which already suggests a reduction in the number of commuter commuters cyclists, but I think the graph showing proportion by time of day below further reinforces that. The proportion of cyclists traveling in the weekday AM peak is much lower after the lockdown, including the sites above that didn't show a large drop in weekday number. There is a clear difference at all sites and in the case of Koonung Trail, it no longer has an AM peak.

With the morning peaks lower most sites showed the cycling numbers increase from mid-morning through to about 6pm. While there appears as a large evening peak in some cases, the lack of morning peak traffic suggests that this is largely recreational riding after school/work unless commuters are spreading out their arrival at work but not their departure.

So what does that mean?

Well it means that cycling patterns have changed. It looks like less people are commuting to work but there are more recreational or non-commuter transportation cyclists out and about. Organisations like Neighbourlytics identified that people are staying more local during the lockdown which may make cycling as transport to shops and cafes a more viable option. Personally I had to stop cycling to the supermarket because by reducing my shopping frequency it meant I couldn't carry everything home!

What does this mean for a return to work?

Well that's harder to predict but I think the data indicates that we haven't seen an increase in commuter cycling traffic, yet. Vehicle traffic in the city is almost back to pre-lockdown levels but pedestrian traffic suggests we are nowhere near the pre-lockdown level of people in the CBD. Cycling provides an excellent commute option for people living close to the city and providing better cycling infrastructure (here and here), like bike lane and trails, has been shown to help more people jump on their bike. Making the change from a car trip will free up road space, or from public transport will fee up physical distancing space which will help avoid COVID-19 transmission and be good for the whole community.

So while overall cycling levels may be down, I think there's potential for an increase in commuter cycling on the horizon as more people return to office work in the CBD. There's an oppotunity for a step change in commuter behavior, by building on the current interest in recreational cycling with better cycling infrastructure. At the very least I hope that's what happens or our road infrastructure might not cope.

Sources:

• Department of Transport - Bicycle volume and speed data