When good data makes bad plots

Intermediate Data Visualization with ggplot2

Rick Scavetta

Founder, Scavetta Academy

Bad plots: style

• Color
  • Not color-blind-friendly (e.g. primarily red and green)
  • Wrong palette for data type (remember sequential, qualitative and divergent)
  • Indistinguishable groups (i.e. colors are too similar)
  • Ugly (high saturation primary colors)
• Text
  • Illegible (e.g. too small, poor resolution)
  • Non-descriptive (e.g. "length" -- of what? which units?)
  • Missing
  • Inappropriate (e.g. comic sans)

Bad plots: structure and content

• Information content
  • Too much information (TMI)
  • Too little information (TLI)
  • No clear message or purpose
• Axes
  • Poor aspect ratio
  • Suppression of the origin
  • Broken x or y axes
  • Common, but unaligned scales
  • Wrong or no transformation

• Statistics
  • Visualization doesn't match actual statistics
• Geometries
  • Wrong plot type
  • Wrong orientation
• Non-data Ink
  • Inappropriate use
• 3D plots
  • Perceptual problems
  • Useless 3rd axis

Wrong orientation

Broken y-axes

Broken y-axes, replace with transformed data

Broken y-axes, use facets

3D plots, without data on the 3rd axis

3D plots, with data on the 3rd axis

Double y-axes

Double y-axis for transformations

Guidelines not rules

• Use your common sense:
  • Is there anything on my plot that obscure a clear reading of the data or the take-home message?

Let's practice!

Intermediate Data Visualization with ggplot2