4  The Method

The following 7-step method is a useful framework for building a visualization that answers a specific question.

1. Make a list of your constraints

   1. Who is your audience? Is this graphic for you, for a report, or for publishing?
   2. Do you have a color scheme or font you are required to use for branding?
   3. What graph types are compatible with your dataset? Is there a field standard?

2. Define a few research questions, then narrow it down to one.

3. Clean and transform your dataset, if needed.

4. Perform an exploratory data analysis and identify which data answers your questions.

   1. This includes statistical summaries, rough visualizations, and filtering.

5. Start with an ugly graph and play around with scales, faceting, and graph types.

6. Fine tune your graphic to maximize the data-ink ratio, fit to your constraints, and make it look nice.

   1. Consider the color, font, and positioning. Do they softly draw your attention to the relevant data?
   2. Add short meaningful labels that do not over explain.
   3. Layer any additional augmentations that color, font, position, and labels don’t address.

7. Repeat steps 3-7 until you are satisfied, but remember that like art or writing, you can always improve.