Data visualizations and infographics have grown in popularity. More and more data is produced and our ability to absorb it in a meaningful manner has decreased.  By relying more on graphics and less on numbers alone to tell the story, the key message(s) can more easily be distilled. (Follow us on Facebook where we frequently share our favorite infographics.)

Cartography takes the infographic concept and applies it to maps in order to illustrate data by geographical region.  In recent months I’ve seen an increasing number of these graphics make their rounds.

Cartography generally takes one of three forms:

  • A choropleth map, which uses shading and/or patterns to show data.  It usually has a “clean” look and is easy to see the region boundaries.  However, incorporate too many colors/patterns and it becomes an effort to read.  The size of any region may mislead, as well.  Despite a color/pattern indicating otherwise, a large region doesn’t necessarily correlate with the data (e.g. a big  region doesn’t equal a larger impact).

via Wikipedia

  • An overlay map is one in which, you guessed it, the data is presented on top of the geographical map.  You can see both the true geography as well as the relative size differences in the data.

    National Election Bubble Map 2008

    via StatChat

 

  • A third type, and actual inspiration for this particular blog post, is the proportional map.  In these maps, the emphasis is on the relative relationships between the regions, rather than accurately depicting geographical boundaries. These maps can really catch the eye and vividly display relationships. Some are practically art in themselves. However, they can also mislead. A primary challenge in interpreting these maps is that the reader first needs to know the underlying  true geography in order to make the most sense of what they are seeing (the example below actually provides the base map at the bottom).

via stephabegg.com

 

All three have their relative strengths and weaknesses. While I love the beauty of the proportional maps, I personally prefer the overlap map in most cases, or, if the data is simple enough, the choropleth map. What do you think? What makes the clearest sense to you?

To view other proportional maps on a variety of subjects, visit http://www.ghemawat.com/maps/.