The Grammar of Graphics

Recall: A statistical graphic is a mapping of variables in a

• data set to
• aes()thetic attributies of
• geom_etric objects.

All pages/chapters refer to the “A Layered Grammar of Graphics” paper.

More components (page 8). From hardest conceptually to easiest:

• scales what axes and legends
• stat statistical transformations of the data (p.11)
• position minor tweeks to the position of elements
• facet how to break up data into subsets and display broken down plots
• coord coordinate system: cartesian, polar, etc.
• Extras titles, axes labels, themes

Understanding the defaults (hence Chapter 4 in the paper). In particular

• Each geom_ has a defined set of aesthetics. Ex: Look at geom_bar on cheatsheet.
• Each geom_ has a default stat. Ex: geom_bar's default stat is count. If you want to use a different stat, like identity, you have to set stat=identity.

• Help files. Ex ?geom_line
• Reread the paper after a few lectures.
• Online help files
• ggplot2 book is under Non-Public section of class GitHub page, code from book here. If you want to read it (optional), I suggest reading
  • Chapter 1,
  • Chapter 3, but it will be hard to grasp the first time
  • Chapter 4
  • Revisit Chapter 3.
  • Chapters 5-10 go into specific details.