Perception

Many scientists equate good design with beauty and are surprisingly easily persuaded by aesthetic qualities. In reality, good design is a measure of usability. Think form follows function — use dictates appearance1. Beauty is a laudable goal, but first-and-formost data visualizations should work.

  • 1 No amount of good design can compensate for poor data quality or analysis. Design is too often used to divert the reader’s attention away from in inadequate or faulty analysis. This workshop will help you to identify cases where visualisations are flawed or misleading, while at the same time helping you make outstanding visualisations. We will begin by considering composition and color.

    • For explanatory data visualisation, this means clear, meaningful and honest communication. For exploratory data visualization, this means understanding your data thoroughly, including diagnostic plots to assess your models, see Table 66.1.

    Let’s begin by considering visual preception. There are two ends of the spectrum, as described in Table 1.

    Slow Fast
    Exploratory phase Explanatory phase
    Confusing Intuitive
    Table look-up Gestalt principles
    Labour intensive Saliency
    Many messages Self-explanatory
    Table 1: The two extremes of visual perception.