Visualizing History of Science
The other day, I was looking through the book Atlas of Science: Visualizing What We Know (MIT Press, 2010) by Katy Börner, a fascinating big-data look at science for historians. It gave me the idea of playing around with visualizations of my own using the Isis Bibliography data. What I’ve produced here are quite unsophisticated by comparison, but they raise interesting questions nonetheless.
What you see in the accompanying images are a few pie charts that show the disciplinary strengths of the Isis Bibliography based on data from 2002 to 2013. They arise out of the subject classification scheme that I set up when I began working on the bibliography. You can find that scheme in its entirety here.
I am tempted to talk about what these charts tell us, but I find that I can’t do that yet. Instead, I will focus here on what questions these charts encourage us to ask. (I highlight this difference because I believe visualizations can be very useful in the discovery process, inspiring us to develop new research agendas. This is a point that often gets forgotten when people see a graph as a QED moment, rather than the depiction of a pattern to be investigated.)
So, let’s think about some of those questions. To what extent are these pie charts telling us about the bibliography itself? It should be clear that part of what we see in these charts is an artifact of the quirks of my classification scheme. Part of it reflects my own choices about whether a work falls into one category or another. And part of it arises out of where my assistants and I have looked for and found publications to include. It is important to consider these factors in order to be able to properly interpret the charts, but they don’t take us very far.
When we take those bibliographical constraints into account, I suspect that we’ll be able to see much more. In particular, what are the charts telling us about the way that the discipline of history of science is structured? By looking at the strengths and weaknesses in the distribution of research areas, we can begin to think about why some subjects are more popular than others.
The most interesting questions, in my view, concern the network of historians working today. What are the institutions supporting historical research and how have they helped to shape our scholarly productions? How have particularly productive scholars and specific research programs shaped the picture? These charts lead us to ask questions that will force us to dig down into the data, that will take us to specific people and institutions. Suddenly, our own discipline’s history begins to open up before us.
Finally, what do these charts tell us about the distribution of resources among the sciences? In other words, what can we learn about scientific practice itself by taking a big-picture view of patterns in the work of historians of science? It must say something. After all, historical research doesn’t get done unless there are archives, journals, and records produced and preserved by scientists.
Trying to disentangle all of these questions is a major research task, but one, it seems, that we should begin working on if we want to better understand what it is we do and if we want to learn how to do it better.