Chapter 5 Avoiding unethical design practices

5.1 Due Diligence for Analysis

• If your visualization is making a claim, you need to have investigated other explanations of the patterns, have a reasonable certainty that the pattern is meaningful, trust/verify that the data were corrected appropriately and are appropriate for the analysis you’re conducting

5.2 Promote Dignity and Agency

• Nothing about us without us
• Choosing labels and colors
• Provide sufficient information for users to make their own meaning

5.3 Reduce Distortion

5.3.1 Color

5.3.2 Shapes

• If shape changes in both dimensions (width and height), make sure data is represented correctly by area, not just one dimension

5.3.3 Truncated Axes

5.4 Compare Like Things

5.4.1 Dual Axis

5.4.2 Normalizing Raw Data

5.5 Be True to the Data

5.5.1 Don’t Cherry-Pick Data

5.5.2 Matching Data With Chart Type

5.5.3 Beware of Conflicts of Interest

5.5.4 Check if subsets exhibit different patterns (Simpson’s Paradox)

5.5.5 Careful Binning

5.6 Proper Citation and Documentation

5.6.1 Transparent Practices