Cities are one of the many phenomena that follow a long-tailed distribution. In simple terms there are a few big cities and lots of small ones. The classic way of showing a long tailed distribution (and the method from which the name is taken) is to produce as plot such as that below:
The infographic at the top of the post by chinfographics.com demonstrates the distribution in a more engaging and constructive way.
One method I have used in the past to demonstrate data with a long tailed distribution is the excellent Wordle tool. I have created a Wordle (below) for the same data (downloaded from Chinfographics). Whilst it does not compete with the Chinfographics infographic in terms of quality, I still think Wordles provide a very simple, and effective, method of displaying data with a “long tail”.
I recently attended the “Summer School in Practical Survey Analysis” hosted by Oxford Unversity’s Department of Sociology. One session was devoted to examples of good data visualization. The example used to demonstrate a good map is shown immediately below and is taken from this page on the Office for National Statistics website.
Many of us felt it would actually serve as an example of poor visualization for reasons that I think are worth mentioning here. Aside from the fact that it is missing a North Arrow and Scale, the map is misleading. The highest category represents areas where between 6.4% to 60.6% of the population are non-white. This range of values grouped together is too large in this context. Additional categories in the data would highlight the exceptional areas and prevent areas with a population of only 7% non-white, for example, being grouped with areas characterised by a population that is over 60% non-white. In addition the spatial units (Administrative Districts) used to plot the data are too course. The Admin. Districts create an impression that the non-white ethnic groups are fairly evenly spread across Central and Southern England, and, aside from London, not clustered in major urban centres as the associated text suggests (ONS, 2004). Reproducing this map with smaller spatial units (such as Super Output Areas) and a greater number of categories for the data would produce a much more accurate picture of the spatial distribution of the non-white population in the UK today. This example demonstrates the importance of a critical eye when viewing all visualizations, no matter how official the publication may be!
