Visualize This: The FlowingData Guide to Design, Visualization, and Statistics

Visual Cues

In Chapter 1, “Telling Stories with Data,” you saw how encodings work. Basically, you have data, and that data is encoded by geometry, color, or animation. Readers then decode those shapes, shades, and movement, mapping them back to numbers. This is the foundation of visualization. Encoding is a visual translation. Decoding helps you see data from a different angle and find patterns that you otherwise would not have seen if you looked only at the data in a table or a spreadsheet.

These encodings are usually straightforward because they are based on mathematical rules. Longer bars represent higher values, and smaller circles represent smaller values. Although your computer makes a lot of decisions during this process, it’s still up to you to pick encodings appropriate for the dataset at hand.

Through all the examples in previous chapters, you’ve seen how good design not only lends to aesthetics, but also makes graphics easier to read and can change how readers actually feel about the data or the story you tell. Graphics with default settings from R or Excel feel raw and mechanical. This isn’t necessarily a bad thing. Maybe that’s all you want to show for an academic report. Or if your graphic is just a supplement to a more important body of writing, it could be better to not detract from what you want people to focus on. Figure 9-7 shows a generic bar plot that is about as plain as plain can be.

If, however, you do want to display your graphic prominently, a quick color ...