Why is there such a big gap between thematic maps and label maps? Both types of maps show data about places. Thematic maps typically use lots of color to show data about places; whereas label maps use a lot of labels to indicate the names of places — plus they use typographic formats such as bold, italics, caps and so on – to show extra data about places. Compare these two US maps (both from the National Atlas of the United States of America 1,2):
On the left, counties are color coded, indicating one data attribute per county (and tiny counties may not be visible). On the right, cities are indicated with text, plus the population is indicated by text size, plus italics are used to indicate if the city is the capitol of a state or country.
Obviously, both maps serve different purposes, but in both cases additional data about places is getting encoded into visual attributes. The difference between thematic and labelled maps is entrenched in our thinking about maps. In cartography textbooks (e.g. Tyner, Brewer, etc), thematic maps are discussed in completely different chapters than labels: text labels aren’t considered to be thematic.
Perhaps it is useful to look back in time to figure out where this split first occurred to provide some insight. Thematic maps have been around for a very long time. Here’s a pair of maps from the 1850’s. On the left is a thematic map by Minard (link) with circle sizes indicating shipments per port. On the right is a contemporary map from an atlas by Heinrich Keipert (link) where city labels indicate information via text, font size, bold, underline and capitalization.
The first Choropleth Map
The earliest choropleth maps (according to Michael Friendly), are from Charles Dupin in 1819 (almost 200 years ago!) with an example shown on the left below (link). Simple grey shading applied across almost equally sized regions makes for a great image showing a broad dark band across the center of France. Again, on the right is a contemporary map, this time by Carey and Buchon (link) and again, this map has variation in typography such as spacing, capitalization, italics and size.
Crome’s Neue Carte von Europa
So where did Dupin get the idea for a thematic map? A big influence on Dupin was the German researcher August Crome. Below is Crome’s Neue Carte von Europa from 1782. This maps shows where various commodities are produced across Europe.
You can see that Crome starts with a base map that has the labeling conventions of the time, for example italics for rivers, all caps for country name, and color to denote country boundaries. Then he adds on top of this map all the content related to his thematic investigation: different kinds of commodities. He displays these as symbols and codes (only a small portion of the legend and map is shown – original map is here).
Base Map Pain
However, Crome can’t differentiate these symbols and codes from the base map using color, font size, case, and italics — because those have already been used in the base map. Even if he were to use them, they wouldn’t stand out because those formats would just be confused with the base map use of those same attributes.
Anyone who’s designed a map knows the pain of base maps: it’s really hard to make your data standout when the basemap is an already noisy and colorful Internet map or satellite image. And, when you’re designing a thematic map, it’s nice to have patterns in your data visually pop-out. So Crome is backed into a corner and uses different symbols for commodities as well as pairs of letters. However, all of these effectively require perception of different shapes, and different shapes don’t visually pop-out (i.e. shape is not preattentive, e.g. scholarpedia visual search, or Bertin).
So Crome’s proto-thematic map was highly popular but there are no patterns that you see – you have to inspect it closely and read all the labels. Instead, Dupin starts with a much simpler base map – outlines of regions – and his dataset is simpler too – just a single variable. As a result, he is able to use an attribute such as brightness or color. He adds labels too, but his labels are simple plain text and the labels are easily skipped by other later map makers.
Could Dupin have used text and typographic formats instead, like the other contemporary label-based maps of the time? It’s an interesting hypothetical question. Bold type has strong preattentive properties (e.g. Strobelt et al). Dupin might not have known about or had access to bold type: it was invented around the same time as his map on the other side of the English channel (1820s). And the first bold-faces were not available in a range of different weights which Dupin would have needed. Similarly, italics of varying slopes, or different styles of underlines wouldn’t have been available to him. As a result, Dupin and his engraver use intensity, which was available to them, and launching the split between thematic maps and label maps.
Here’s a thematic map, using font weight (more examples of typographic thematic maps are in the paper just published for ICC available here):
I wonder what Crome and Dupin would have thought?