Hey folks,

As I mentioned last week, I’m considering jumping platforms to use Substack to disseminate these newsletters and try to build more interactivity between you all and me. Because of how different email providers process emails, it’s become very difficult to have any confidence in how many people read these emails and who all is interacting with the content. Since I pay by the subscriber I’m likely paying for subscribers who may never open the emails (although their email provider makes it look like they do). Substack is free for me and there are options to keep it free for you or to eventually charge a small amount. Can you let me know if you use Substack to read and comment on newsletters? I figure if people are already using Substack, they might be more likely to engage.

This week, I came across a wonderful scene from the book Underworld by Don DeLillo that relates to the type of thing I’m trying to do with this newsletter and my YouTube content. I admit that I haven’t read the actual book, but this link provides the text of the scene.

There are two characters in the scene. I gather that one is a mentor to the other. The mentor asks the student to name all the parts of a shoe. Maybe pause for a moment and think through the parts on your own. I came up with sole, tongue, laces, heel. The character came up with laces, sole, and heel. With some help he got tongue.

If you’re like me, you thought, “I know there are more parts to a shoe! Why can’t I think of them?” The mentor scolds the student and is able to extract the “front” and “top” from him. The student gets for his inelegant jargon. The mentor goes on to point to the cuff, counter, quarter, welt, vamp, eyelet, aglet, grommet, and the last. In the midst of all this, he makes this profound observation…

Let’s take a step back and grab a plot. Here’s one for you that was recently published in Nature Cell Biology.

Take five minutes and make a list of the different parts of this panel. How many did you come up with?

In my four-step approach to critique - description, analysis, interpretation, and judgement, the analysis stage engages in this exercise. What are the parts? In the grammar of graphics which R’s {ggplot} is based on, there are the data, geometries, aesthetics, scales, transformations, statistics, annotations, and themes. These parts have subparts. The aesthetics include the x and y axis, color, shape, size, transparency and more. The y-axis may be on a categorical or continuous scale. It might be transformed by a linear or logarithmic function.

As the mentor chided the student, we don’t know how to look at plots because we don’t know the parts. Well, hopefully we’re all learning the parts! But that seminar speaker with the pie chart whose wedges added up to 200% doesn’t :) If we know the parts then we can talk about them with precision. Jargon is often off-putting because it excludes outsiders. But for insiders, it is invaluable because they can be precise in what they’re talking about. I would have to look up a picture of a “welt” to know what it is. But I’m grateful it’s there to help keep my boots waterproof. Similarly, one may not know what the expansion is on a plot’s axes, but you’ll be grateful for how it gives some air around the data in the plot.

Let’s move on one step forward building from the Underworld scene. If you have your list of parts from the plot I gave you, can you tell me what they’re “made of”? For example, what variables are mapped to each axis? Which variable is mapped to the color or the shape or the size of the point? Are the axes continuous or categorical? Are the axes transformed? What order are the categorical variables presented and why?

I suspect the mentor from Underworld thought that “seeing” the parts would enable the student to understand the shoe and the person who wore it. The student might have said the sole was rubber, the laces were cotton, and the red uppers were made of canvas. You might have a pair of Chuck Taylors in your head and think they belonged to a kid or a hipster. Alternatively, if we knew the toe had steel in it, we might think it is a work boot of a tradesman. If the materials included silk, we might assume it is a ballerina’s shoe.

Alternatively, we might reframe the comment

Similar to how we might make associations between the parts and materials used to construct someone’s footwear, effective data visualization requires that we know the parts of a plot and what types of “materials” we can use for each part and what materials work well together. We might make choices to make the story more accessible to our audience (e.g., expert vs. outsider, publication vs. slide deck). Sometimes the choices lead to an incorrect interpretation like zooming in on the y-axis when you are using a bar plot.

Let’s close by returning to my four-step process of critiquing a plot. The analysis phase allows us to name the parts and identify the materials that are used for each part. These parts and materials need to fit with the context, which is established in the description phase. In the interpretation phase we integrate the parts, materials, and context to see the story. Similar to how we could assign ownership of different types of shoes to people based on the parts and materials of the shoes. If the context, parts, and materials are not aligned with each other, the interpretation will be challenging. In the judgement phase, we highlight what we like about the combination of parts and materials and try to come up with alternatives that will do a better job of telling the story.

The first step is knowing the names.

​

​

​