Thursday, 28 July 2011

Illustrative maps in current affairs

[Update: I noted on many of my Google Fusion Table posts that, while the data are still on Google Drive for you to view, GFT no longer offers a polygon or heatmap option, only geocoding by country centroid in its new version. Not sure why, but on this, thisthis and another example posted as Iframes not Scripts preserved the old GFT maps.]

The Data section of British paper The Guardian is a great example of illustrating reams of data and helping readers make sense of it - such maps are only illustrations, not exacting science as in my previous post - readers wish to grasp trends for tabular data by country, rather than examine their exact geographies.

Google Fusion Tables (GFT) are an easy way to post and combine data stored by country, say, from the UN Data repository. I already wrote up using World Bank Ease of doing Business data, as well as my mash-up of WolframAlpha search results for religions per capita. Let's now use World Bank data from the GFT Catalogue to do the following:
  1. Find the World Bank 2007 GDP data
  2. Merge it with the World Population over 10M searched in Wolfram Alpha
  3. GFT allows to very easilly merge them:
    a) by Country into a GDP by Population over 10M table
    b) and to add Religions to the same table
  4. Download them locally to calculate per Capita columns
  5. Finally upload them as:
    a) GDP over 10M Population per Capita
    b) GDP for the same by Religion.
But what does all this look like? Well try it out yourself! For the table in section 3.a) above select Visualise then Intensity Map, and then in turn pick GDP as posted below. Then from item 5.a) per capita, pick GDPpCAP as posted below that:

As noted in my earlier post, appearances can be deceptive, as Islam cut a sharper profile in North Africa and South Asia than when it is factored by Population. Notice here how GDP per Capita here is more spread out to lesser populated countries.

So per Capita seems to be a truer measure as it filters out population distortion. But where it gets really interesting is comparing GDP by Religions per Capita. Try it yourself again, go to 5.b) and select per Capita GDP for Islam (IslamPCAP) and Judaism (JudaPCAP), respectively at top and bottom here:

The dichotomy in the distribution not only of population but also of wealth is quite apparent in these simple maps. It appears like the Islamic and Jewish worlds have very little overlap, except in France and perhaps in England. You can study this further with the Christian, Hindu or Buddhist populations. Are these maps therefore not illustrative what may not otherwise be self-evident in current affairs? But equally important, was it not relatively easy to show this once you knew how and where to fetch the data?
As my gran said: learn something every day, and you'll stay young... Go ahead, try it for yourself with other data in the World Bank catalogue!
A few closing notes:
  • There was also a glitch in the data - I had to spell out US and UK in one table to be able to join it to another, and likewise with China and China(PRC) - then again it's the map that made the mismatch readily apparent
  • GFT works for illustration, but it is simple (the legend is fixed), map projections aren't ideal (Mercator projection exaggerate the polar regions), and GFT embeddable code pins maps at the upper left hand corner
    Therefore make sure you scroll the map down, not only to see more of the populated areas and the legend in the lower left hand corner, but also to minimise the distortion in what you see
  • And a GIS function, not feasible in GFT-as-illustration, would be to calculate the geographic overlap - I say: some apps are better for some maps...