This post is the first of a blog series on historical data vis titled “Data Trails. Snapshots from the history of data visualisation” and originally appeared over at my friends’ from Idalab, a Berlin-based specialist on data science and machine learning.
81 years of budget data and various categories in three diagrams – the United States Fiscal Chart from the 1870 US census atlas is a real blockbuster in the history of data visualisation. The atlas as a whole is full of interesting graphics and has a widespread reputation as an early gem of data visualisation.
Today, experts from the field (such as Google’s Fernanda Viégas in this recent panel on the past and future of AI) like to joke that data visualisation serves as a gateway drug to statistics, drawing on our immense visual ability to spot patterns and outliers. Data visualisation is now widely acknowledged as a secret weapon for educating the public about all kinds of interesting problems.This has not always been a familiar approach, though. In fact, among statisticians there has always been a strand of resistance against graphics, assuming that beautiful visuals might give readers a wrong impression: „Statistics should be the driest of all reading“, as William Farr, a 19 th century British epidemiologist once put it.
Francis Amasa Walker, on the other hand, was not afraid of colourful graphics. Born in Boston, he was appointed superintendent of the 9th US census in 1870. Apparently he was a man of many interests, and it seems not just a little coincidence that he had previously worked as a college teacher and as a newspaper editor – both of which are jobs in which he routinely had to consider precisely how to present information. So when he and his team set out to publish the data from the 9th census in an atlas, they went to great lenghts to create attractive and accessible visualisations and maps.
Thus is the context for this awesome fiscal chart, which never ceases to amaze me with its 81 years of US budget data. It covers the period from 1789 to 1870. Imagine: Back in 1789, when the US were born, data tables had to be written by hand and kept in the drawer. It seems quite an achievement to me that those early budget records were still completely available when Francis Amasa Walker set out to create this chart some 80 years later.
The chart forms a vertical timeline, starting with 1789 at the top and running all the way down to 1870. The left diagram presents the structure of revenues, showing that the United States during this period generated much of their public income through customs. The diagram on the right shows the structure of public spending, again differentiated by categories. The central column is taken by a diagram of public debt. Revenues and expenditures are broken down by category, but not compared by their relative amount. The debt chart, on the other hand, encodes the total amount of debt in the width of each line.
The chart is a prime example of the aforementioned powers of data visualisation: to show patterns and outliers. What becomes immediately present is an overall change in the US budget’s structure during the years 1861 through 1867. Most of the expenditures go to the army, while a large chunk of the revenues is now generated from „internal revenues“. The debt chart in the middle would immediately explain what that is: all of a sudden, US public debt explodes. That is quite a striking method to demonstrate how much of an (economic) desaster the American Civil War has been for the United States.
Image via Library of Congress, where the digitised atlas can be viewed in full (the fiscal chart is image 86 of 107).