Design Exercise 1-2: Why This Visualization

in Design Exercises (List Page)

In this second design exercise, we’ll look at some visualizations and try to figure out why the designer chose the specific designs. The goal is to build intuitions about task and effectiveness in visualizations. You will turn in your answers to questions about the visualizations on Canvas as Design Exercise 1-2: Why These Charts? (due Fri, Sep 12).

Introduction

This assignment asks you to look at a paper (about visualization) - not necessarily for the content of the paper, but rather, to look at the visualizations and try to understand why the author chose the visualizations that they did.

I’ve chosen a paper that Connor Bailey and I have written recently - because it has 4 different (but similar) visualization designs within the span of a short (4 page) paper. (4 designs in addition to the 2 visualization designs being tested in the experiment!). The paper is available here (direct to pdf).

We are not asking you to read the paper. We are asking you to answer questions about the visualizations in it. You might need to look at it a little to see why the visualizations are there (to get some context). But, the points of the paper are not relevant to the class at this point.

For 3 of the visualizations (figure 2A, figure 2B, and figure 3) I would like you to answer:

  1. What do you think the visualization is trying to make easy to see? In other words… What is the visualization trying to convey? What are the key tasks / communicative goals?
  2. What does the design do to make these things easier to see? (or, more broadly, why do you think the authors chose this design? what are some things that lead to its effectiveness?)

There are 6 questions on the Canvas survey (Design Exercise 1-2). (2 for each of the 3 visualizations)

Note: I am not asking you for something else the visualization could have shown (but doesn’t), or for a critique.

These 3 different visualizations have very different goals, and therefore very different designs. Hopefully, it will help you appreciate why we need to adapt designs to achieve specific tasks.

Another note: at this point, we haven’t learned much about “visualization science” - you might not be able to describe or identify what is going on. It would be interesting to try this again later in the semester. But for now, use your eyes and intuitions.

An Example

I didn’t ask you to try this for Figure 1B in the paper… Because I am going to do it myself here as an example.

Figure 1A is just an illustration of the test conditions of the paper. It is a visualization, but it has a much simpler goal.

What do you think the visualization is trying to make easy to see? This visualization is trying to make the general trend in the results easy to see: that over the 4 conditions error and response time decrease between the different conditions. The visualization also makes it easy to see that the trends between the two different chart types are very similar. Close examination reveals the confidence intervals of each measurement.

How does the design make these things easier to see? The design put the measurement on a common axis and connects the related measurements with lines, so the trend is visible. The rows (conditions) are sorted from worst to best, so the trend is simple and stands out. Each axis is next to the corresponding example stimulus allowing so the viewer know what each row means.

(not a question for this assignment) It would be easy to critique how this could be better - for example, it is not easy to see the distinctness in the confidence intervals (that they don’t overlap), or to read precise values. If the reader cares about this, they need to read deeper into the paper.

Mechanics

Since this is one of the first assignments, I’ll review the mechanics…

This assignment is part of Module 1, and is due at the end of the module. We’ll do a similar exercise in class, so you might want to wait and do this closer to the due date.

You will turn this assignment in as a Canvas Survey Design Exercise 1-2: Why These Charts? (due Fri, Sep 12). When you turn it in, Canvas will give you 87 (a middle AB) points automatically. A human grader may give you more points if your answer is particularly good, or take away points if your answers aren’t worthy of an AB.

The assignment is due at the end of the module. Late assignments are accepted according to the late policy (see Syllabus (late policy)).