Tuesday, 5 March 2013

Maps for England compendium


An interesting compendium of maps was recently published by a group of researchers in the UK. They extracted a whole slew of maps from various government reports and web sites or from compiled statistics to support a report to the Royal Town and Planning Institute (RTPI). The report itself also includes some additional maps.  It has a secondary use... as a resource for geography teachers with an almost infinite choice of examples to illustrate sub-par mapping. There’s about 80 maps, only a handful of which are well designed.

Rather than providing a systematic critique (and sending readers to sleep)...here are some general observations. See how many maps you can find that fall into these traps...and which ones clearly deal well with these issues:

  • Gross regional representations of data which will have very little policy relevance at a local level (ecological fallacy)
  • Poorly generalised boundary data (i.e. used as is and inappropriate at that scale)
  • Poorly edited boundary data (spurious polygons purporting to be administrative boundaries, often in coastal areas and including large bodies of water)
  • Poor use of map space and contextual detail (Wales and Scotland could be displayed to avoid England appearing to be an island...Northern Ireland often shown without the Republic of Ireland)
  • Poor titles (often including acronyms that need explaining)
  • Legends that display class breaks to n decimal places (round up)        
  • Lack of labels (makes it difficult to interpret)
  • Poor mix of basemaps and thematic detail (topographic basemaps clutter the map display and reduce legibility of thematic overlay)
  • Use of highly saturated colours (they just don’t look good!)
  • Heavy linework (over-emphasises least important features)
  • Sparse map space with detail only in highly urbanised areas (so just show local maps)
  • Overlapping classes (some areas could fall in more than one class and it’s not clear which they do fall in)
  • Poorly scaled proportional symbols (creates a flat map of data that looks similar but isn’t)
  • Unintelligable wording (gross-linkage?)
  • No need for a map at all (“Great Britain shown for illustrative purposes only”...so why a map? use a graph!)
  • Gawdy diverging and dichromatic colour schemes for sequential data that has no critical value (use a single hue sequential scheme)
  • Overlapping dots (it masks much of the detail in crowded areas)
  • Random unique colours used for sequential data (and black as one of them!)
  • Class breaks that contain gaps (suggests you should use a different scheme)
  • Poor contrast between the important mapped theme and other background content (make your data tell the story...not be part of the background)
  • Data overload (too much content for one map to handle with overlapping symbology prevailing)
  • Black labels that conflict with black linework (makes them unreadable)
  • Choropleths with totals (most avoid this pitfall...but not all; data has to be relative)
  • Poorly structured data (just because layers stack in your GIS, you need to use sensible drawing levels to make them work on the map)
  • How can sites of active and proposed nuclear power plants possibly cover the extent they do?
  • Symbols so small it’s impossible to see their colour (which is the only differentiating visual variable used)


Finally...ask yourself a single question when you’re looking at any of the maps...is the map as it stands capable of properly supporting policy-making? I struggled with most of them because while many of the above issues might be deemed cartographic gloss, without the care that goes along with the gloss the evidence simply isn’t compelling. It’s weak...but worryingly all too often taken at face value and used in anger...and that’s when the real mistakes are made!

2 comments: