Learning from the Past

I think there is a lot to learn from our past, especially in a subject like data visualization that feels like it’s moving so fast right now. So much of the more recent work is variations and adaptions of what has come before, that it’s useful to understand these earlier ideas in context.

I’m currently reading Tukey’s 1977 Exploratory Data Analysis. My Dad gave me a copy of this book when I first moved to California years ago, and had good foresight that I might find it interesting.

Early in the discussion of plotting, Tukey takes a few paragraphs to describe materials. He writes “If we want to see what our plots ought to tell us, there is no substitute for the use of tracing paper (or acetate). If we slip a well-printed sheet of graph paper just below the top sheet of a pad of tracing paper, we can plot on that top sheet of tracing paper almost as easily as if it were itself ruled. Then when we have the points plotted, some boundary or reference lines drawn, and a few scale points ticked, we can take away the graph sheet and look at the points undisturbed by a grid. We often gain noticeably in insight by doing this. (And, we have had to pay for a sheet of tracing paper rather than for a sheet of graph paper).”

Moreover, despite the challenge, expense, and manual work of plotting numbers by drawing them, Tukey continues later in the chapter that: “We almost always want to look at the numbers. We do not always have graph paper at hand. There is no excuse for failing to plot and look.” [emphasis in the original text].

It’s interesting to see the empahsis on the importance of graphing, to see the data visually in order to understand the data, even when graphing was so cumbersome.