General interest in data visualization has been steadily growing over the past few years.
The digital age and in particular the availability of data has been pushing us towards better ways to convey messages hidden behind numbers and tables.
According to google trends, interest in data visualization has grown 250% over the past 8 years.
But how did this general interest in #dataviz look when there was no internet, no TV, no radio, or cars?
That's a different story.
And we will tell you part of it from the perspective of data visualizations themselves. In those pre-computer times you could not simply input an array of variables and wait for the magic to happen on a screen. Alternative methods were used to store and put information in context so one could visually retrieve it. Here are some amazing examples of physical visualizations and why we should thank those patient visualizers.
One of the very first visualizations known to mankind came from South America and were actually talking knots, or Quipus from their original in Quechua. Quipus were recording devices that encoded numbers (in a base ten positional system) using knots in strings and categorical information using colors. They were one of the methods of population calculation for the Incas. The oldest known Quipu is 4600 years old. By the time of Conquistadores they were still in use but soon suppressed because they were considered idolatry by Roman Catholics.
Quipus are a long lasting visual artifact used to encode information.
Polynesian Stick Charts
Living in 1,000 islands in the central and southern Pacific Ocean makes you develop a different sense on how to move around, especially in the times when Polynesians used their memory to navigate without any complex mechanical aids.
Their only aid were Stick Charts.
Stick Charts were used to navigate the Marshall Islands' coasts. Islands were represented by shells tied to the sticks, or by knots in two or more sticks. Stick charts would also be used to encode information about currents, waves and ocean swells.
Physical Viz Meet Statistics
Charles Davenport was an American eugenicist and biologist that used to explain different concepts using arrangements of physical objects. Today Davenport's research seems controversial, but he was indeed a pioneer in biology working, among other things, in new quantitative standards of taxonomy.
In 1901 Davenport "built" physical visualizations that show the distributions of features of objects and people. His purpose was to explain the notion of statistical distribution to a lay audience.
His visual explanations are of great pedagogical value.
The image show seashells piled up according to how many ribs they have, we might call this a proto histogram.
This image shows a group of students from the University of Chicago. On the left they are sorted by their height, on the right they are arranged in files by their height group. This top view of the students is critical to represent every student as a point independent of their height, as we are interested in seeing the observation counts of the different height groupings.
Willard Cope Brinton was a consulting engineer and a pioneer in data visualization. In 1914 he published "Graphic methods for presenting facts" and devoted a whole chapter to maps and pins: 30 pages that cover probably everything that could be known at the time about this visual representation.
Different mechanisms for encoding data were used in these physical maps, from using different pin shapes and colors (left), to making physical zooms by extracting highly dense areas out of the map (see the New York area cluster floating off the coast to the right).
Working with physical visualizations requires a few extra considerations, in the case of pin maps you might want to follow Brinton's suggestion of using corrugated straw-board. For a more in depth review of pin maps see this great post.
Cosmographs, Flow Charts and Sankey Diagrams
A cosmograph sounds like complicated physicist machinery, but in fact it is a visualization technique created by an Irish Engineer Matthew Sanke according to Jim Strickland (2012). It also happened that cosmographs are explained in Brinton's book, in there he explains that these graphs are particularly useful for representing flows, like those in financial inputs and outputs.
Source: Images adapted from http://dataphys.org/list/ibms-cosmograph/
In the image we can see a cosmograph built by IBM around 1933. Users would group its 1000 strips of paper in groups to depict flow changes. The left image shows how the paper is glued together to represent flows as percentages. The interesting fact is that cosmographs were actually meant to be photographed rather than read directly, that's why they could have repeatable and clean reproductions like the one in the right of the image.
These types of visualization are today more commonly known as Sankey Diagrams, they are also sometimes called flow charts but the name is too generic to stick.
These examples of physical visualizations along many more were the starting point for today's explosion in data visualization. They paved the way to hundreds of ways we can represent data visually nowadays.
The best collection on the web on physical data visualizations is http://dataphys.org/list/. We thank the contributors of this site for their amazing work and invite you all to discover some gems of data visualization history and interesting facts.