Scales: Defining How Data is Mapped to Aesthetics

🧠 Map ’n’ Shape

When you specify an aesthetic like x = wday or color = cyl, you’re telling ggplot2 what variable to map to the respective aesthetic — andggplot2 encodes them accordingly.

But what defines how the values of that column are translated into visual properties? That’s where scales come in.

Each aesthetic (like color, size, or x) has a corresponding scale that controls how the data is visually represented.

Together, aesthetics and scales form a two-part system:

• aes() maps your data columns to visual properties.
• scale_*() controls how those properties appear in the final plot.

The names of scale functions (and there are a ton!) always reference the respective aesthetic and often a specific type of data or behavior:

scale_[aesthetic]_[type]()

Feels abstract? Let’s look at a few examples:

• scale_x_date()
  
  Controls how date variables are represented along the x-axis.
  
  
• scale_y_continuous()
  
  Controls how continuous (numeric) variables are displayed on the y-axis.
  
  
• scale_color_discrete()
  
  Controls the encoding of discrete (categorical) variables mapped to the color aesthetic.
  
  

Based on the mapping you’ve specified, ggplot2 applies appropriate default scales (if you don't specify them yourself) — behind the scenes, to make your plot work:

🎚️ Controlling Scales

If you want to tweak how a scale behaves, you can simply modify its settings — or replace it with the scale_*() function that fits your needs.

Some common reasons to modify or replace a scale:

• Changing axis limits — like filtering the data range or capping extreme values
• Switching color palettes — like using custom colors for discrete groups or gradients for continuous data
• Styling axes and legends — like adjusting tick marks, re-labeling categories, or modifying the legend’s appearance
• Transforming the data — like applying a log scale on the y-axis or aggregating values into bins

You don’t have to change the data to do any of this — it’s all handled by the scale_*() functions.

Most scale_*() functions have share some general arguments:

• name: set a custom axis or legend title
• breaks: define steps to display along the axis or in the legend
• labels: customize the value text on the axis or in a legend