Like everything else, the dynamics of the Glasgow bike share scheme have been significantly altered over the course of 2020 due to the ongoing pandemic. Though I covered the scheme in some detail in March, the situation has changed to an extent that another look at the data is warranted.

In this post, I'd like to examine the changes to the scheme using data collected here at SchemeStats. If you're interested in the data presented, feel free to check it out for yourselves using the dedicated Dashboard, Live Map, Metrics Map, and Day Profiles tools. I hope you find it useful.

Contents

• Pre-Pandemic Review
• 2020 vs 2019: A Significant Increase in Demand
• Changing Day Profiles
• Station Flows
• A Spectacular Saturday
• Conclusions

Pre-Pandemic Review

The earlier, pre-Covid post on Glasgow came to the conclusions that:

• The scheme experiences a daily usage increase each weekday morning and evening, with peak usage at 9 AM and 5:30 PM respectively. This indicates that the scheme was being used for commuting purposes.
• Peak weekday utilisation figures were typically around the 6% mark, so at its busiest one in sixteen bikes were in use.
• The commuting periods saw a drainage of bikes from outer areas in the morning and an increase in bikes in the city centre, and the converse each evening.
• Weekend activity was more subdued than weekdays, with a gradual increase from late morning throughout the afternoon. Sundays were busier than Saturdays.
• Availability of bikes was good, with most stations having bikes available most of the time. There appeared to be additional capacity in the scheme for additonal usage without compromising bike availability. 

2020 vs 2019: A Significant Increase in Demand

SchemeStats does not collect data on rentals directly, so we need an indirect way of measuring usage. The quantity of bikes in transit is already estimated based off the quantity of bikes in stations at any given point in time, so it is a short hop to extrapolate the amount of time spent cycling from that figure. If we take it that every minute one bike is in circulation is one minute of cycling time, then the amount of cycling time per day is simply the summation of the minute-by-minute quantities of bikes in circulation.

Using this approach, it is possible to get an indication of usage month-on-month, and that's what I've done below to compare usage in 2020 vs 2019. Note that data for July 2020 only runs up to the 29th, so the final figure for the month will have been slightly higher.

*July's figures only up unto the 29th.

Clearly, usage is significantly up on the same period from last year. It is interesting that the gap in usage really takes hold at the same time as the pandemic unfolded. It is to be expected that usage would rise over the summer months with improved weather conditions, but the divergence in figures for April to June in particular illustrates how different a year 2020 has been for the scheme.

Changing Day Profiles - Much Busier Evenings

A ‘Day Profile’ is a plot of the average values of a given metric on a given day of the week – e.g. ‘Available Bikes on Tuesdays, January to March’. Below, I've used this tool to create a time lapse of week-by-week profiles of two key metrics: Bikes in Circulation and Empty stations. I hope this will give us a better understanding of how the scheme is changing over time. The figures here are for the first week in January until the week ending July 26th.

Bikes in Circulation

Let's start by looking at the Bikes in Circulation (BIC) figure for this year. Note that the range below doesn't suit the figures for the first few weeks of the year as it's far too large a range on the Y-Axis, but this is necessary because of the dramit increase in activity later weeks see. To get a better view of what the BIC profile used to look like, take a look at the earlier post. There are a couple of points in which there appear to be radical dips in Saturday's usage late in the animation but please excuse this error in data processing - the adjacent figures are correct.

Some observations from the changing BIC profiles:

• From mid March the morning commuting peak disappears, correlating with when initial measures society took to mitigate the effects of the pandemic took place.
• Since mid March we see some radically busy weekends, both on Saturday and Sunday. Figures are variable, but the sample sizes are small and over the course of these weeks combined it appears safe to say weekend activity overall has picked up.
• Any evening commuting activity has been subsumed since the pandemic started into a much larger, gradual increase and decrease in activity each weekday, starting around midday and not winding down until nightfall.

But might the pattern we're seeing above not just be typical as the weather gets better? Well, no - for two reasons:

• We know overall usage is up dramatically after looking at the cycling time statistics at the outset.
• Last year's day profiles for BIC don't look like those above. Let's take a look...

Here's the same chart for the same weeks in 2019. Unfortunately the data is a little patchy as there were a couple of bugs last year, but there is still enough to illustrate how much quieter the scheme was last year. The scale is the same as above.

One observation to take away from last year's figures is that the weekday commuting pattern of morning and evening peaks didn't really manifest itself until May of last year. This is in contrast to the profiles from this year where such a pattern was already evident in January. So it appears that to start with, the scheme was already more widely used this year than last.

Empty Stations

This dramatic increase in BIC figures above must be challenging the availability of bikes. This is clearly illustrated below by an animation of the evolving day profiles of the number of empty stations in the scheme.

Some observations:

• Availability is very favourable until April.
• From April onwards, the baseline figure of empty stations rises from below 5 stations to 10-15 stations .
• The increased activity in the afternoons noted above is resulting in a corresponding sharp rise in empty stations over the same periods of the day.
• Some weekends have been particulary acute, with over half the stations being empty for a portion of those days.

Station Flows

Using SchemeStats’ Metrics Map tool, it’s possible to plot the geography of a given station measure over a given time period. That time period can either be a simple from- and to- value, or a slice of a day. We can plot things like ‘The average net change of all stations between the hours of 7AM and 10AM, Monday to Friday’. Using this tool, I've plotted the change in net flow figures for every stations during the morning and evening traditional commuting periods. As above, each frame in the animation is one week of activity. Data is for Monday to Friday only. Red stations are gaining bikes, and blue stations are losing them.

Morning Commutes - Weakened but Returning

Here is the evolution of net change in bikes per station from 7 AM to 10 AM, Monday to Friday, week on week:

In the early weeks of the year, the traditional morning commuting pattern from outer areas to the centre was clearly in evidence. However we see clearly that from mid March, activity cools completely. However, in the subsequent weeks there is a notable increase in activity, albeit not to the same levels as before.

Evening Commutes

Here is the evolution of net change in bikes per station from 4 PM to 7 PM, Monday to Friday, week on week:

Observations:

• Again, activity dropped off in mid March
• Activity returned rapidly, unlike the morning commute which is still recovering
• The pattern is different from the original evening commute - activity is more dispersed.
• We know that there are large numbers of bikes in circulation in the evenings from April onwards, but this isn't manifesting in as clear a pattern as we might expect if activity was entirely driven by commuting.

A Spectacular Saturday

Before coming to the conclusions, I'd like to Illustrate just how busy Glasgow's scheme can get lately. Weekends in particular have shown a shift to remarkable uptake during the afternoons. As Insider.co.uk reported recently, Saturday June 20th saw record-breaking usage, so that seemed a good day to examine. To illustrate this day, I've animated the Live Map and the Dashboard for Glasgow from 5AM to midnight. Note that utilisation of bikes during the afternoon reached a remarkable 80%!

Conclusions

• Glasgow's bike share scheme is in use far more than last year.
• Morning commuting activity has dropped off significantly but is recovering since the effects of the pandemic began to take hold.
• The evening commuting period has been replaced by a gradual increase in activity over the course of weekday afternoons.
• Weekday afternoon activity is more disperse than before, suggesting use for leisure rather than commuting.
• Weekends can be extremely busy, though variable, with huge spikes in usage during the afternoons. Utilisation during these times can be well over 50% of all available bikes, with Saturday June 20th seeing utilisation over 80% percent.