Visualization Best Practices

Know your audience

Think about the purpose of your visualization. Are you explaining or exploring a phenomenon? Will your audience better understand and respond to a simple respresentation of data, or something more complex.

Provide context and incorporate instructions

The typical academic research paper provides a story arc, with introduction, background, literature review, methodology, discussion, and conclusion, which Stephanie Evergreen effectively points out in Effective Data Visualization, may not be appropriate for all audiences. In "fast-paced, decision-making contexts," she writes, "I don't think we actually want a story. We want an interpretation." This nuance is especially important when visualizing data. If your audience needs to reach a decision using available data, help your audience interpret the data you have visualized. Use graph titles and other visualization techniques to help audience members focus their attention towards key takeaways in your data.

Add interactivity

Add interactivity to your visualization if possible, using tools like R Shiny or Tableau. This can allow anyone using your visualization to apply filters or seek aditional context for a data point.

Apply visual best practices

Choose an effective visual

Choosing an effective visual is one of the most difficult and rewarding aspects of visualizing data. Fortunately, a small number of graphs, when skillfully formatted and presented, fulfill most data visualization needs.

If you wish to display a relationship, comparison, composition, or distribution, Abela's Chart Chooser directs you to various charts based on the number of dimensions -- or categories -- and/or the number of measures -- or variables --- you have to visualize.

Quantitative/Qualitative Chart Chooser

If you are thinking about how to influence action using the data you've collected, Stephanie Evergreen's Quantitative Chart Chooser, which is published on the inside front cover of her book Effective Data Visualization, encourages you to consider what "you need the audience to do when viewing the data." If you need your audience to focus on a single number, try a large call-out number or icon array. If you need to show how a number changes over time, try a line chart or slope graph.

If you are struggling to find an effective way to share qualitative data, Evergreen's Qualitative Chart Chooser, which is published on the inside back cover, helps to connect the story in your data with effective visualization types. (See Chapter 8: "When Words Have Meaning: Visualizing Qualitative Data."

What chart to choose depends on your audience, your message - or what you're trying to show - and the dimensions and measures in your dataset.

Test, test, test

Ask a colleague or friend to critique your visual. Try visualizing your data in two or three different ways to determine which presentation best suits your audience.