A History of Data Visualization up to 1900
A look at the origins of the field.
Kris Sankaran (UW Madison)
01-07-2024
Reading, Recording, Rmarkdown
- There could probably be an entire class taught on the history of data
visualization. The reason it is worth covering in an otherwise
practically-oriented class is that,
- Historical study can illuminate the core intellectual foundation on which the
entire discipline of data visualization is built.
- The limits, biases, and trends of our current era become clear when
considering it within its full historical context.
- It’s often possible to draw inspiration from past masters.
- It can be humbling to realize that now-commonplace ideas had to be discovered.
If it took until 1833 until the first scatterplot was published, then what ideas
have we yet to find?
- The reading divides up the history of data visualization into 8 epochs. In
these notes, we will consider the 5 epochs before 1900.
Before 1600: Early Maps and Diagrams
- It might come as a surprise, but visualization has been around since the
invention of writing. The Egyptians made maps, and there are examples multiple
time series plots from the 10th century.
- However, most visualizations were focused on physical geographical or
astronomical quantities, and even these quantities were only imprecisely
measured. Without more formal data gathering instruments, there could not be
much data visualization.
1600 - 1700: Measurement and Theory
- The situation begins to change around 1600, at the dawn of the scientific
revolution. All of a sudden, precise measurement of physical space become
possible. Also, important new mathematical ideas were introduced, like
probability, calculus, and reasoning about functions. This created the right
environment for the design of some of the first truly sophisticated data
visualizations, like
- Christopher Scheiner’s 1630 visualization of sunspots over
time (a first instance of faceting),
- Edmond Halley’s plot of barometric pressure against altitude (an first
instance of one feature being plotted against another), and
- Edmond Halley’s plot of wind speed over the ocean (a first visualization of a
vector field).
- During this period, the scientific value of visual thinking was more or less
established. However, there were still only relatively few graphical forms
available, and the focus continued to remain on visualizing physical quantities,
rather than more general social, economic, biological, or ecological data.
1700 - 1800
During this period, many new graphical forms were invented, including
timelines, cartograms, functional interpolations, line graphs, bar charts, and
pie charts. Further, forms from earlier, like maps and function plots, became
more firmly established.
This was also when three-color printing was invented. Before this point,
color could not be used as an encoding channel.
In this century, governments also began large-scale data collection of social
and economic statistics. One of
the most prolific inventors of data visualizations, William Playfair, was used
visualization to study a variety of economic problems. In addition to inventing
line, bar, and pie charts, he experimented with original ways of composing
multiple graphs to suggest the relationships between variables.
1800 - 1850
This was a period of maturation for the field of data visualization. By this
point, visualization had become standard in scientific publications. Advances in
printing technology also made it also became easier to mass produce
visualizations.
It was also around this time that the scope of problems studied through
visualization expanded far beyond display of purely physical (geographical and
astronomical) applications. Two areas in particular flourished, applications to
social science and to engineering.
- Visualization in the social sciences began to emerged in response to
government sponsored collection of social statistics – data about crime,
births, and deaths, among other topics. This wasn’t a purely intellectual
exercise: understanding demographic trends was important for countries that
were often at war with another.
- In engineering, the idea that visualization could serve as a computational
aid become more and more common. For example, rather the chart below, by Charles
Joseph Minard, displays the cost of transporting goods across different
stretches of a canal. Vertical breaks correspond to cities along the canal, and
the area of the square between cities encodes the cost of transportation between
those cities. While the information could be stored in a table, it becomes
easier to perform mental computations (and make guesstimates) using the display.
1850 - 1900: The Golden Age
It might be counterintuitive that there was a golden age of visualization a
century before the first computers were invented. However, a look at the
visualizations from this period demonstrate that this was a period where
visualizations inspired scientific discoveries, informed commercial decisions,
and guided social reform.
For example, in public health, Florence Nightingale invented new
visualizations to demonstrate the impact of sanitary practices in
hospital-induced infections and death. Similarly, it was a
visualization that guided John
Snow to the source of the 1855 cholera epidemic.
- Some of the graphical innovations include,
- 3D function plots. The plot below, by Luigi Perozzo, shows population size
broken down into age groups and traced over time.
- Flow diagrams. Charles-Joseph Minard, who we met before with the canal
visualization, was a master of these displays. One in particular is widely
considered a masterpiece, it shows the size of Napoleon’s army during it’s
Russian Campaign.
- Multivariate visualization. Francis Galton made some of the first efforts to
visualize more than 3 variables at a time. His Meteorographica, published in
1863, contained over 600 visualizations of weather data that had been collected
for decades, but never visualized. One plot, shown below, led to the discovery
of anticyclones.
- This was also an age of state-sponsored atlases. More than sponsoring the
collection of data, governments assembled teams to visualize the results for
official publication. From 1879 to 1897, the French Ministry of Public Works
published the Albums de Statistique Graphique, which under the guidance of
Émile Cheysson, developed some of the most imaginative and ambitious
visualizations of the era.
Citation
For attribution, please cite this work as
Sankaran (2024, Jan. 7). STAT 436 (Spring 2024): A History of Data Visualization up to 1900. Retrieved from https://krisrs1128.github.io/stat436_s24/website/stat436_s24/posts/2024-12-27-week14-4/
BibTeX citation
@misc{sankaran2024a,
author = {Sankaran, Kris},
title = {STAT 436 (Spring 2024): A History of Data Visualization up to 1900},
url = {https://krisrs1128.github.io/stat436_s24/website/stat436_s24/posts/2024-12-27-week14-4/},
year = {2024}
}