I have long wanted to post about the many ‘news stories’ on the web, usually spread via footer advertising, which make a determined story out of quite storyless data. The ones where the author has decided a story before they have even seen the data. The simplest example is where a top 10 list has been compiled in the hope that it will uncover some interesting trend but doesn’t. I recently I stumbled across a good example, a top 100 that while factual I believe makes unfair conclusions from the underlying geographic data.
The article I stumbled across is very typical in that its topic is somewhat unrelated to the site on which it was found. The article in quesion is on the UK website of The Weather Channel.
Here is the article, entitled ‘The 100 fattest towns in the UK’. It is based upon a story/interview from The Telegraph and refers to a geo-statistical dataset that was complied by Public Health England, the medical and scientific report of which was published in the journal The Lancet.
Whilst most of my disgruntles are with the reporting of location-based facts, I have to start by pointing out two somewhat misleading errors in the title. Firsly the dataset covers England, not the whole of the United Kingdom (UK). Secondly, the captured data appears to be by something like unitary authority, council region, output area or NHS trust, i.e. the administrative areas of England, and so the majority of the ‘places’ in the dataset are not actually towns.
To prove why this is a non-story, I decided to produce a map of the twenty most overweight “towns” in the “UK”.
Not only does it show the points raised above but did you notice that I have sized the symbols proportional to the percentage of obesity in each area? No, I can’t really see the difference in size either. I purposely omitted value labels so you could see how the variation in symbol size is not appreciable. The reason being that in the data there is actually very little difference in the percentage of population overweight in the highest ranked location, Copeland (75.9%), to the location ranked 50th, Wakefield (69.6%)! As a science graduate I find it almost unethical to report these results in such a definitive way, where is error accounted for? The only conclusion we can sensibly draw from the data is that far too many people in England are overweight and it is a problem that is spread right across the country.
The Guardian tried to map similar data back in 2008. I have to say that their attempt is just as bad if not worse. Apart from possibly an attack on Wales by fat people, what exactly does it show?
Both maps highlight the point that to make maps properly, a cartographer should always refer to the original data source whenever possible and display the data appropriately.
The use of points is misleading as the data refers to geographic areas. As cartographers, we often make urban settlements (towns and cities) points on maps at smaller scales, but to be true to the data, which in this case is given as percentages and based upon vastly different geographic extents, then a thematic map of the administrative polygons would probably be better. The BBC for example used a correct method to display the 2008 data, although perhaps not with the best or nicest of colour palettes:
The quandry here is that the correct map is not headline grabbing, so despite everything I have said I can completely understand why media would take a route more like the Daily Mail below:
It would be possible of course to use the correct method yet still create a attractive map from the data. Does anyone know any good cartographic design consultants?