Gordon Stokes Transport

Mapping 2011 census journey to work flows

Transport Home

Photography Home

Transport Photos

Last updated - Sept 2017

More on ...

Pages in this series

This page

Interactive census travel mapping

The material on this page was mainly created in 2015. I've recently added some interactive mapping where you can select some or all modes and look at different areas by zooming and panning(see the "Interactive census travel mapping" link to the left, though you might want to read this page first before looking at a local area).

Census flows introduction

These pages show mapping of travel to work flows from the 2011 census. They are based on data from Neighbourhood Statistics and NOMIS, and use Quantum GIS (QGIS) for mapping, making special use of the "FlowMapper" plug in. I've improved on this mapping (see above).


The 2011 England and Wales census was based on peoples' usual residence and also asked for the address of their main place of work. It also asked the usual 'method of travel' (mode) used to get to work. In 2013 the Office of National Statistics published the method of travel for residents across its different geographies (scales of unit for amalgamating data), and in 2014 NOMIS published the flows between areas by different modes, at Local Authority and Medium level Super Output Area (MSOA) level.

Matrices of travel to work journeys were downloaded from NOMIS. There is a limit of 1,000,000 records per download which means travel between 1000 zones is the maximum that can be gained in one download. All regions except South East England have fewer than 1000 MSOAs, so one mode at a time can be downloaded for all other entire regions. But 'cross-boundary' journeys won't be shown, so overlaps were incorporated - for example, for the South West (with about 700 MSOAs) I added neighbouring Local Authorities in Hampshire, Wiltshire, Worcestershire and other places. This isn't perfect, as there will be journeys (escpecially by rail) that will be missed, but I reckon that all major flows by modes other than rail are all shown.

Because the MSOAs have very roughly similar populations the flows are broadly indicative of the relative scale of travel. But because urban areas have a higher population density this can distort the appearance of travel intensity. A small town is likely to have one zone that includes most of its employment so radial patterns will emerge, but a larger settlement with more dispersed employment may show a more diffuse pattern. In a place with homes and jobs dispersed around a wide area, flows may get 'diluted' - the appearance probably exaggerates the 'suburb to centre' flows. Journeys within one zone aren't shown on these maps but can be seen on the interactive mapping pages.

The data was prepared for use with "flowmapper" which requires geographic co-ordinates for the zones (I used population weighted centroids downloaded from the Office of National Statistics [ONS]), and a matrix of flows between each of the zones. These were mapped using QGIS, and flows banded such that:

Why am I doing this?

This mapping is of dubious direct relevance for transport planning since its accuracy is nowhere near perfect. But a prominent transport analyst did say "anyone working in placemaking ought to read it".

My interest is in the differences between places, and how this level of visualisation shows that simple rules such as settlement size or gravity modelling do not explain many very significant differences that exist between apparently similar towns or cities. As far as I know schools no longer teach that Luton is where they make hats, or Witney where they make blankets, but these maps show cases where a modelling approach may miss out on important local historic or cultural differences. Issues such as cycling levels in Milton Keynes and Stevenage also point to the danger of a 'build it and they will come' way of thinking (see the cycling text).

So I put these here for others to look at and suggest ideas for why things are as they are, and hopefully improve on the methodology and interpretation. Also, as a part time artist (photographer) I think some make interesting images.

Example maps

Various maps are shown here, and most are similar in form to the one shown below, though most are not as complex. This map shows much of South East England with all main modes in different colours. While the first map has place names I haven't for the others, but the interactive maps do have placenames. On these you'll have to use another map if you don't know which town is where.

Motorways are shown as curving blue lines, and A roads as thin red lines. B roads are also shown in brown. It should be quite easy to guess which mode is which, but the key is below. I have also overlain modes on top of each other - also explained below.

More detailed mapping can be found on the "Interactive census travel mapping" page (at the top of this page)..


The lowest (furthest back) layer is rail, followed by metro, car driver, bus, walk and then cycle on top. This ordering was used so that, in general, each mode shows up. But is some places such as Oxford cycling virtually obliterates walking and bus and car.


The following maps show different modes separately. The first is for walking, and it's pretty easy to see that most urban areas have a lot of walk to work trips over a short distance. Besides London, the amount of walking seems to be pretty much in relation to the size of city/ town.




Cycling shows a different picture alltogether. Here the places that stand out are Oxford, Cambridge and parts of London, especially to the North East and South West of the centre. But unlike walking the patterns in other towns don't seem very predictable at all. Remember that these are just journeys to work, so Oxford and Cambridge would be much busier if students were shown. Bedford and St Neots, East of Cambridge, and Abingdon, near Oxford show up, but most large towns such as Reading and Milton Keynes hardly show at all. This is especially astonishing for Milton Keynes which has one of the most highly developed cycle route networks of anywhere in the UK, even Europe. The same is true of Stevenage and some other post war new towns, built at a time when cars had panache and were the future (until such time as we all had jet-packs), while cycling tended to be regarded as an inferior means of travel for low paid workers in cloth caps - not what new town residents wanted!




Bus use also varies more than would be expected by urban size alone. And bus is generally only widely used for shortish trips. The bus catchment for central London is a similar distance to many other towns and cities (though of course London has a very well developed metro and rail system which is more convenient for longer journeys). Many London suburban centres also have their distinctive catchments, the most pronounced of which is Heathrow. Romford, Croydon, Kingston upon Thames and many others also feature as separately functioning bus centres. It looks like buses haven't done much for Canary Wharf.

Outside London, Reading and Oxford are most pronounced. Reading has a long history of high bus use, while Oxford was one place that actually benefitted from deregulation competition in terms of increased bus use rather than bus operators pricing each other out of the market. The recent complete removal of bus subsidies in Oxfordshire aren't so impressive. Northampton, Luton and Cambridge also show. Oxford, and to a lesser extent, Cambridge also have longer distance bus journeys, reflecting a scatter which will be visible when car driver patterns are looked at.


Note that I changed A roads to green on this map so that the red shows up.


Buses tend to work best when timetables can allow a bus to complete its route in just under half an hour. This means that timetabling is easier with a regular timetable and getting more journeys out of one bus than if it took, say 35 minutes. Also, most people don't like to be on a bus for more than about 20 minutes. This helps explain why buses tend to be well used for fairly short distances, and also why some towns and cities work better or worse in terms of bus travel. Reading and Leicester used to be held up as examples of 'good bus operating territory' and this is probably because buses could get from one side to the other in a easy timetabled time. Leicester is much bigger than Reading in terms of population but is also much more densely populated.

It's also apparent from the map below of England and Wales that the conurbations also show bus use patterns that revolve around the sub-centres that most of us who don't live there tend to forget about. Bradford, Huddersfield, Dewsbury, and Wakefield (West Yorkshire); Barnsley, Rotherham and Doncaster (South Yorkshire); Oldham, Bolton, Hyde, Bury, and Manchester Airport (Greater Manchester), and Wolverhampton, Dudley, Coventry, West Bromwich and Stourbridge (West Midlands) all have their own distinctine bus catchments.

There are a few places where the radial out from a town pattern doesn't hold up. Without any heavy bus corridors the area around Bangor and Caernarvon in North Wales has a small spider's web. There are similar things a few other places. What I'm not so sure about is the long lines between Workington/ Maryport and Appleby in Cumbria. I'd assumed I'd got the zones mixed up but repeated it and got the same result. It could be bussing of workers, or a lot of people taking the micky on their census form (or, possibly, I've still got it wrong). But I did later find some very odd cases while doing the interactive mapping and ONS confirmed that there are errors in the data. I think that's fair enough, as I doubt I could ever produce a perfect set of data at this scale!



Rail and metro

This is pretty much self explanatory. London dominates, and the Underground (metro) is better developed to the north of London. Non London radial routes are hard to spot, though Reading, Didcot, Oxford and Banbury can be seen to be linked by modest flows.

But some rail corridors into London are much better developed than others. Cambridge has far greater flows than its equivalent of Oxford, which could be due to the frequent coach service from Oxford, and may also relate to train frequency and continued use of elderly shaky dumpy little three-car noisy cramped but otherwise 'delightful' trains that have been known to run pretty much on time (sometimes) which carry people from Oxford to London for many peak hour trains. Having said that the 125 trains are still probably the most comfortable way to spend half an hour sitting outside Reading Station - but I digress. Oh but I should add that at some point in the very distant future Oxford will get smart new electric trains that will flash by the over-engineered gantries that are being put up in the Areas of Outstanding Natural Beauty. And since December 2016 trains have operated from Oxford station to Marylebone which are more modern and seem less susceptible to delays. By 2021 the electric trains still won't be running but plenty of people from London will have found the sky-high housing prices around Oxford cheap enough to live in Oxford and commute using the Marylebone line either from Oxford or from the Park and Ride just to the North.



Car driver

Driving is the most heavily used mode of transport for work (around 60% of journeys to work are driven by car). A map of the same format is shown as for the other modes above, and below is a smaller scale national map for England and Wales. A few comments from the many that could be made on the first map:



The national map shows (amongst other things) that:



For more detailed mapping go to the Interactive Maps - link at top.


Back to Transport Index

Gordon Stokes, 2016