Depicting Human Issues with Human Cartograms

Cartography, to the best of my knowledge, is the study and method of map making. It is distinct from mapping or geography in that the latter two generally use maps to analyse regions or present findings. Typically, a standard land area map is used, with points or areas of interest highlighted.

Cartography, on the other hand, is interested in things like how the map is projected from a three dimensional shape (the Earth) on to a two-dimensional piece of paper. The choice of projection method – and creating new methods of projection – are the domain of the cartographer. Again, that's to the best of my knowledge, so don't quote me!

Human cartography, then, is the study and method of producing maps that best highlight human issues, in contrast to a traditional approach which uses a geographical area map to which additional information is somehow added or superimposed. Perhaps I had better illustrate this with an example.

The following map is taken from who have kindly licensed their maps under a Creative Commons licence. The map is a standard land area map of the world; no surprises:

I'm not sure of the exact projection method used – probably Mercator – but it's a standard map that we all recognise. The main limitation of this projection is that land areas closer to the equator are depicted as smaller than they really are, whereas areas closer to the poles are depicted as bigger than they normally are. Nevertheless, if this map were shaded or coloured with a theme such as life expectancy, you would know how to read it.

But a human cartogram is different. It produces a map that is still recognisable (if you recognise the original, like the world map), but it distorts the map to illustrate a human topic. For example, the following map – again kindly taken from – depicts the prevalence of HIV/AIDS (that is, how many people have HIV/AIDS):

Now, I knew that many countries in Africa, as well as India, had the highest prevalence of HIV and AIDS in the world, but this map is a striking illustration of the extent of both HIV/AIDS and it's unequal distribution among countries.

Producing Your Own Cartogram and other sources have produced a wealth of human cartograms free to use and analyse, and these are an amazing resources. However, they typically only show information on the global level, i.e. the whole world, like the maps above.

Human cartograms can, however, be produced for any geographical region, provided you have data (or at least estimated data) for that region. You need a shapefile, which is a file produced by GIS software and has an extension .shp. The shape file needs to contain information about the geography or region you are studying, as well as the data on your human topic (like HIV/AIDS prevalence or life expectancy).

Sadly, the easiest way to produce a shape file is with proprietary software like ArcMap. If you have access to such software through your institution or organisation, great. I encourage you to use it. If you don't, there are open source (and free) alternatives available such as QGIS and I've even seen R used for GIS. I know nothing yet about using them, but am booked on a course later this year to learn so I'm sure there will be a subsequent blog post about open source GIS soon.

The source you use for your geographical boundaries will of course vary depending on your region or area of interest. I know, for example, that the UK Data Service provide census boundary data for all regions in the UK, which is what I have used to construct my cartograms. Once you have obtained your boundary data this can be imported in to ArcMap (or alternative).

Next, you should join your human topic data to your geographical data in your GIS software, such as ArcMap. There are various data sources available, and I have again been using the UK Data Service website to obtain data. There are two important things to remember when obtaining data:

  • You must have some way to link your data to the geographical boundaries. This is normally done by a code (for example a ward code) and is like a database key so the ArcMap knows where that data is for._


  • Your data must be a count. You cannot use means, percentage, rates, etc. The acid test is this: you must be able to add your data up and for that total to be a useful statistic. So, using the example above, you can add up the number of HIV/AIDS patients in each country, but you would not add up the percentage of HIV/AIDS patients in each country because that would be meaningless (instead you would probably work out an average percentage).

Once you have your geographical boundary data and have joined your data set, you need to export this as your shape file. You can then use a free piece of software called ScapeToad to produce your cartogram. The process of producing your cartogram is straightforward (at least compared to obtaining and saving your data in the necessary format!). ScapeToad is java-based, so runs on any and all operating systems with java available.

To produce your cartogram, load your new shape file. Tell ScapeToad where to find the boundary data and then which is the thematic data. You can then produce your own cartogram for your own region of interest.

An Example Cartogram: Owner Occupancy in England

I've produced an example below. Both maps depict the number of people by unitary authority or district in England who do not own their own house (or have a mortgage on it). I've used this because it is one of four census themes used in the Townsend Material Deprivation Index, so it is one way to indicate relative deprivation.

The first map is a traditional geographical map of England that is instantly recognisable. Each district or unitary authority is shaded to highlight the number of people in that area who do not own their home. It's crude, but the darker the shade of grey, the more people who do not own their own, and arguably the greater deprivation. The map is roughly what I would expect to see, and overall this map does not have any great surprises. There is low owner-occupancy in central London, for example.

In contrast, the human cartogram below is still recognisable as England but it is much easier to immediately determine the areas that have low owner-occupancy because they appear larger.

There are clearly a large number of people who do not own their own home in central London, and to a lesser extent areas in the midlands, the north west and the north east. Conversely, comparatively more people own their own homes in Devon, Cornwall, Norfolk and in the Cotswolds and areas of central England.

I'm not intending to start a debate about the merits of the Townsend Material Deprivation indices, or the use of home ownership as a measure of deprivation, or even of how to interpret these maps. Instead I just want to illustrate that human cartograms offer a powerful alternative to depict and present geographical data where:

  • Good quality data is available about the geography in question.
  • The audience is, or is likely to be, familiar with the geography discussed (for example England, the World). I think this technique is less successful where the geography is less familiar (for example wards, LSOAs). It is of course possible to provide a standard geographical map of the area beforehand for comparison, but I believe this technique is strongest and most striking when the audience is already highly familiar with the area.
  • A human theme is being depicted (for example poverty or deprivation).

Maps from © Copyright Sasi Group (University of Sheffield) and Mark Newman (University of Michigan).

Some of the data for the example is taken from the 2011 Census. Census output is Crown copyright and is reproduced with the permission of the Controller of HMSO and the Queen's Printer for Scotland.