In today’s world of visualization, the bar chart is one of the most widely used statistical diagrams. William Playfair, an ingenious rogue Scotsman, is often credited with having invented this type of diagram around 1786.
But was he really the first to work with this method? Where did he draw his inspiration from?
William Playfair is one of the most fascinating figures in the history of information visualisation, not just for his “interesting” biography. He invented three types of statistical graphics — seemingly out of nowhere — which today are the bread and butter of our daily viz routines: the line graph, the bar chart, and the pie chart. The former two were first exemplified in his 1786 Commercial and Political Atlas (here’s a good overview of the charts contained in this atlas).
Nineteen of the twenty charts in the book were line graphs, while only the last one (pictured here above) introduced a different innovative technique of representing numbers. It should be interesting to learn why Playfair presented this in a horizontal format, with the bars running from left to right, but this may be due to the horizontal format of the atlas and the other nineteen charts. Either way, I was thrilled to come across two very interesting “proto-bar charts” during my research for The History of Information Graphics. Both pieces preceded Playfair only by a few years, and – interestingly – both covered the same topic.
1000 Years of River Floods along the Elbe
One of them was published in Germany, just two years prior to Playfair’s atlas:
This ambitious data adventure (I mean, go figure: the book presents 1,000 years of flood data!) was published by Christian Gottlieb Poetzsch, a citizen of my hometown Leipzig. He was a self-taught naturalist and collected all kinds of data, e.g. about the soil and the weather. He was also very interested in watching flood levels of the rivers in the region, and was the first person to install a gauge in the Elbe—one of the main rivers in Eastern Germany—near the town of Meissen.
The book, which is otherwise composed merely of text and a few isolated data tables, contains this enormous foldout diagram in the end, and I was thrilled when I first saw it. It shows the scale of the river gauges along the y-axis, and analog to the rising floods, the bars rise from the bottom upwards. Although the time is running from left to right, the x-axis doesn’t show a continuous time period. Instead, each bar represents a distinct event within the given period (1501-1784).
What I find really exciting about this diagram is how the graphic depiction of flood levels in individual bars seems to be an ingenious “transcript” from the actual situation of measuring the water levels along a gauge stick. The gauge stick is the defining item here and, consequentially, it has not only served as an abstract scale reference, but is also depicted visually – which makes it even easier for the reader to understand this piece.
30 Years of Water Levels of the Seine
The other example that I came across covers the exact same topic—flood levels of a major river—and was published by the celebrated French cartographer Philippe Buache in Paris in 1770, 14 years before the Elbe piece above:
Philippe Buache: Eaux de la Seine (1770), sheet with copperplate engraving from Cartes et tables de la geographie physique ou naturelle. Courtesy David Rumsey Map Collection
Buache, too, was a geographer and naturalist with a broad variety of interests. He conducted various geographical, geological and hydrographical studies. He is recognized as a pioneer of thematic cartography, as he continuously searched for methods of graphically representing his findings and the data collected along the way.
The sheet pictured above resulted from his longitudinal studies of the course of the river Seine. Beginning in 1730, he gathered information about the path, seasonal variations and natural obstacles of the river, and he had two markers of river height installed in Paris, one at Pont Royal, and the second at Pont Tournelle, from which he obtained daily (!) measurements.
In this sheet, he compared the water levels measured over three decades, with the Seine flood of 1740 as the maximum level of 25 feet. Another reference line (shaded blue-green) is introduced at appr. 20 feet, bearing the date of dec. 31st, 1769. Unlike in the piece above, the x-axis here does represent a continuous time frame, and every year is represented with four measures: for each half of a year, both the highest and the lowest levels are given. The specific dates and measures can be derived from the data table in the top part.
The sheet by Buache is a little less direct in its visual references to the actual measuring set-up with a gauge stick than we could observe above in the Elbe-piece. However, Buache, too, clearly includes a few visual clues: the striped layout of the gauge sticks is reprised in the baseline as well as in the vertical lines separating the decades. Also, a brilliant feature is the extremely delicate hachure which differentiates the bars — and is clearly a visual reference to water, as that’s what is being discussed here. One bar represents one half year. While the full bar is shaded in a common tone (hand-colored, by the way, using a brush and water color), the lower part of the bar is shaded darker in an engraved hachure.
Making the transfer — from geography to economy
Ok, before I get lost completely marveling at these graphic details let me wrap up what this is all about. The bar chart is one of our absolutely essential basics in information visualisation. I would argue that the format initially (and quite naturally) originated in gauge measurements of river levels. Up to now, we do not know whether these two authors were the only ones to draw such diagrams.
We also don’t know whether or where Playfair had seen those. However, he had contacts in the scientific circles of Paris, where Buache was an important figure. It is not too far fetched to assume Playfair could have had access to Buache’s atlas Cartes et tables de la geographie physique ou naturelle.
It was Playfair’s ingenious achievement to adopt this graphic method—invented to represent hydrographic data—and turn it around to represent abstract economic numbers. This is no small achievement to make. Inspired by the writings of Adam Smith, he was searching for graphic representations for quantitative economic data, and found it in the highly developed scientific cartography. It takes a creative and brilliant mind to bring together intellectual impulses from different scholarly fields, such as thematic cartography and economic research.
But it is also important to understand that Playfair was not alone in the desert. He worked from within a developed culture of scholarly research (in Europe and increasingly in the then emerging US) in which many scholars and authors contributed to developing the graphic methods we are working with today.
Image sources can be accessed by clicking on the images. If anyone comes across another early river level diagram (from before 1786) or any other related material, please send it my way and I will update this article.