The latest instalment of our weekly series, in which we use the Centre for Cities’ data tools to crunch some of the numbers on Britain’s cities.

Over the last few weeks we’ve looked at Britain’s commuting patterns in a variety of ways. We’ve looked at which cities. We’ve looked at which cities cycle most, and which have the busiest public transport. (We haven’t yet covered which cities are most car dependent, but rest assured, we will.)

What we haven’t done, though, is looked at how those stats relate to each other. That’s today’s job.

So, how do those stats relate to each other? They relate to each other like this:

Source: Centre for Cities, based on ONS data.

A brief explanation. The horizontal axis is the proportion of a city’s workers that, as of the 2011 census, commuted to work by car. The vertical axis is the same figure for public transport. And the bubble size represents the number who cycle to work.

What the graph shows is quite how close the inverse correlation between driving and public transport is. More people on public transport (as in London, at the very top) means fewer in cars. More people in cars (as in Swansea and Telford, in the bottom right) means fewer in public transport. The vast majority of cities in Britain are pretty tight to that line.


The only cities that break with this relationship – that have fewer people on both public transport and in cars – tend to be shown with much bigger bubbles. In other words, in Cambridge (the huge bubble in the bottom left) and Oxford (the slightly smaller one above it), so many people cycle to work that it skews the other two figures.

The correlation isn’t perfect, because these three methods of commuting are not the only option. The chart doesn’t show people who walk to work, or who work from home, or who commute by some mysterious “other” method. (Unicycle? Horseback?) Add those in, and you get to 100 per cent.

But all the same, this graph is a helpful illustration of an obvious statistical truth: if you want to get people out of their cars, you have to provide them with an alternative.

Here’s a chart of the same data.