I write a lot about typography and visualization. It all started with critically looking at maps and noticing differences between modern visualization and old maps. I did a PhD looking at typography, text and visualization. (Stay tuned, there will even be a book in late 2020 about visualizing with text – with many new visualizations beyond what I had in my thesis!)
Back to maps. I was invited to speak at ESAD Valence about visualization and I decided to take a break from book writing and revisit the original inspiration: maps. Cartography has different rules than visualization, a much longer history, and many different techniques readily visible. So, I cobbled together some of my favorite maps to talk about and point out some observations.
Gough Map, 1360
The Gough map is a wonderful medieval hand-drawn map. Rivers are diagrammatic starting as bullets and flowing in almost straight lines. The iconography for towns varies from simple sheds, to an added cathedral tower, to a cluster of small buildings, to the walled city of London. Typographically, it’s interesting with an ordering of labels. While most towns are labeled in brown, London is literally labelled in gold. Distances between towns are labelled in red, and counties are labelled in red with boxes (e.g. Suffolk).
Munster’s Geographia Universalis, 1540
Skipping ahead two centuries, Munster’s maps from Geographia Universalis (1540) are interesting maps at the transition to the printing press. Like the medieval Gough map, rivers, mountains and towns are highly stylized forms and pictographs, which are combined together with typographically differentiated text in italics, caps and roman. Although the geographic map is a woodcut, the lettering is highly uniform and likely metal type composed together with the woodcut by a form cutter. The resulting aesthetic balances the rougher shapes and textures of the woodcut with the fine metal letters plus some ingenuity by the artisans to get it all fit together. Towns are consistently horizontal but labels are angled to fit, such as Vincentza turned almost upside down:
Willem Janszoon Blaeu, 1629
Engraving enabled much finer detail than feasible with woodcuts: both the topography and the labels could be engraved in detail. Willem Janszoon Blaeu‘s maps have an expanded set of iconography, now reduced even smaller to tents, pyramids and tiny houses. The path of rivers is more accurate and mountains have shading. The engraved text now has more opportunity for variation. River labels more closely align with river courses. Labels corresponding to areas are larger and spacing starts to increase (e.g. D A N). Plus many other text variants (size, case, italics) differentiate between names of towns, cities, provinces and regions.
Crome’s Neue Carte von Europa, 1782
Crome creates an early thematic map, Neue Carte Von Europa, showing location of different crops, livestock and minerals in Europe in 1782 (previous post). An even wider range of icons are now required to indicate all the different types of resources: gold, silver, copper, zinc, iron, mercury, marble, fruit, honey, salt, rice, fish, wood, horses, pigs, etc. — 56 different types of commodities. After running out of icons, two letter codes are used, e.g. Kr for cork, Tb for tobacco, Cr for currants and so on.
Sherman’s map, 1864
During the U.S. Civil war, general Sherman lead his army deep into the Confederacy, far beyond his supply lines. Sherman’s map combines traditional topographic detail with an overlay of resources summarized from the 1860 census. Starting with a base map showing counties, cities, rivers and railroads, an additional 15 variables of census data are added regarding the quantitative resources available: population, livestock, and agriculture. The map provide Sherman with the ability “to act with confidence that insured success.” As an early datamap for analytical and planning purposes, it shows the value of depicting many dimensions of data simultaneously, to aid in trade-off decisions, such as food available, potential resistance and potential supporters.
Ordnance Survey, 1921
Modern maps, using printing presses, reach a high in the early 20th century for the amount of information packed into them. Ordnance survey are a favorite for the amount of information that they pack into each label. In this example from the early 1920’s, place names vary capitalization, italics, size, font family (plus the actual name) to indicate 5 attributes per label (legend here).
Steiler’s Atlas, 1924
Similar to the Ordnance survey, mapmakers on the continent also created maps with high-dimensional labels. Stieler‘s maps are typographically interesting as the labels use an ordering of underlines (dot, dash, solid, double solid) to indicate cities with different levels of governance (e.g. capital of a county, province or country). Also, backward italics for water features, curved and spaced test to indicate area features, and so on.
FAA Aeronautic Chart, 2019
Here’s a map that’s only a few months old from FAA.gov, and packed with a phenomenal amount of information for pilots. There are many different classes of information, visually distinct from each other. The base map has topographical shading in hilly areas, bright yellow in urban areas. Overlaid are blue and red layers, each with a wealth of information regarding the corresponding airport, runway configuration, airspace, routes, waypoints, radio frequency, visual markers such as stadiums, wide turbines and bridges, and more. Icons and alphanumeric codes are heavily used to compact data for expert users. All text remains legible, with the background/basemap largely being light/bright upon which other layers can be superimposed, and if needed, some text is set with light halos.
Even though most people might think of Google maps these days, with minimal representation of roads and highly undifferentiated labels, the history of maps shows far richer solutions packed with many layers of information. These much richer maps, like the aeronautic chart and Sherman’s map, show that there are uses and applications where people need more information than only a couple classes of information within one visualization. And all the examples here show how all this extra data can be communicated with labels, symbols, lines, layers and more.
So, where and when could scatterplots, timeseries charts and treemaps add many layers to increase their information content and aid new analytical uses?