Last modified: 20 February 2017
There are lots of good reasons to keep the UK capital where it is. I disagree with most of them, but they exist.
There are also lots of good reasons to move the UK capital from London to somewhere further North. There’s the fact that London is so unlike the rest of the UK , or that residential and commercial rents in the city are so high, or that the Houses of Parliament are about to collapse.
But there’s an argument that’s used both to justify London’s central role in the UK and to back making a change. Some people say that London is a centrally-located and accessible city — the natural place for the country to meet. Others say that it’s in a corner of Great Britain, and an especially long way from Northern Ireland, Scotland, and Wales.
What does the data say?
Well, we know that the centre of population of Great Britain is somewhere near Leicester , having moved South from somewhere near Derby a century ago.
But what really matters is how easy it is to get somewhere. I’ve been investigating that.
The proposed 2018 UK parliamentary constituencies have some very useful properties — they cover the whole of the UK, they have an almost-equal population, and their location is well defined.
So I took the point at the centre of the 600 proposed constituencies and asked Bing Maps for directions to The Houses of Parliament, Manchester Town Hall, and Birmingham County Hall. I added York Minster in later, because I can.
This is what that my calculation looks like from Derby North to Birmingham County Hall on a map. I calculate a journey time for driving on the left and public transport on the right and I record the time for both. Then repeat it for every constituency.
Instead of using the Bing Maps website, I use the Bing Maps API to do all of this automatically. It takes about a second. I even wrote a lovely Universal Windows app to help me.
I’ll just dump the results here. There's a spreadsheet too. It’s pretty simple, Birmingham wins. London does better than we’d expect from geography alone — the railways are typically much faster towards London — but it’s still beaten by Birmingham on most measures.
I could write more, but I won’t. For now.
I can already predict some of the replies to this piece. “600 constituencies isn’t nearly enough” [I did calculations against all 7201 MSOAs to check, same result], “you should have used PWCs not simple centres” [the MSOA calculations were PWCs], “you’re not considering traffic”, “you’re not considering public transport service frequency”, “you’re not considering flights” [Bing’s API doesn’t have this feature].
I’m getting increasingly frustrated by replies like this. We are swimming in a sea of data and far too many intelligent people revel in it. They develop hugely complex models and tests that provide no extra value and instead exclude others from understanding or criticising what they’ve done.
I hugely value criticism. I’m even known to pay for it. I dislike dismissing criticism. But if your comment is “you should have done [hugely complex thing that would probably make no difference but make the result much harder to understand]”, I am unlikely to reply. Oh and I calculated the journey times in both directions, I started all public transport journeys at 7am, and I deleted all constituencies where a journey time could not be calculated for any destination. With public transport directions this included the whole of Northern Ireland.