In christyray/crthemes: A Standardized ggplot2 Theme for Visualizations

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 6.5,
  fig.asp = 2/3,
  out.width = "75%",
  dev = "ragg_png"
)

library(ggplot2)
library(crthemes)

Updated r format(Sys.time(), "%b %d, %Y") \ \ The crthemes package includes two functions for modifying the legends of bar graphs and heat maps to better work with the rest of the theme. The examples here show those functions and some ways to use the standard ggplot functions to improve the output of fill-based graphs.

Bar Graphs

The standard bar graph legend has the fill boxes very close together, and there isn't currently a way to space them out further using the standard theme function. The theme functions are applied somewhat unintuitively to the legend spacing due to how the legends are built on grid graphics (source).

To workaround this, we can define a new key-drawing function, as described here and here. The key drawing function in this package is draw_key_polygon2() with the wrapper function key_scale(). They can be used with key_glyph = key_scale() inside a geom function.

The inputs to the draw_key functions are the plot data, additional params that specific to the geom being plotted, and the size of the key space in mm. These are all available in the ggplot2 environment and do not need to be defined by the user.

To increase the space between the legend keys, the draw_key_polygon2() function provides a modifier that scales the legend key down within its allotted space. This gives the appearance of increased space between the keys by making them smaller. The key_scale() wrapper function allows the user to specify a scaling factor, with a default scaling of 0.8.

Examples

The plot code used for the examples is below. The expand1() function is a convenience wrapper for expansion(mult = c(0,0.1)) to remove the space between the bars and the x axis.

df <- data.frame(
  x = factor(c("IL6R", "IL8R", "IL6R-Ab", "IL8R-Ab", "IL6R-Ab-IL8R"),
             levels = c("IL6R", "IL8R", "IL6R-Ab", "IL8R-Ab", "IL6R-Ab-IL8R")),
  y = c(0.45, 0.78, 0.61, 0.31, 0.72))

p <- ggplot(df, aes(x, y, fill = x)) +
  scale_y_continuous(expand = expand1()) +
  scale_fill_cr(palette = "receptors", species = levels(df$x)) +
  labs(x = NULL, y = "Y Axis", title = "Bar Graph", fill = "Species") +
  theme_cr()

\ \ The ggplot default legend spacing:

p + geom_col()

\ \ \ The key_scale() function is used with the key_glyph argument of a geom function. It only has one argument, scale, which defaults to 0.8 (80 percent). The legend spacing using the key_scale() default:

p + geom_col(key_glyph = key_scale()) 

\ \ \ A value for the scaling factor can also be passed to key_scale. Smaller key scaling will lead to more space between the legend keys:

p + geom_col(key_glyph = key_scale(0.5))

\ \ \ Using the legend.key.size argument of the theme() function increases the size of the space allotted to the legend keys. This will add more space between the text lines and will make the keys look larger even when they are scaled down:

p + geom_col(key_glyph = key_scale(0.5)) + theme(legend.key.size = unit(24, "pt"))

Heat Maps

The color bar accompanying heat maps and other plots with a continuous fill scale is typically significantly shorter than the height of the plot. The heatmap_legend() function uses the guide_colourbar() function to increase the length of the color bar and scale it to the size of the plot. It also removes lines for the upper and lower limits and scales the size of the ticks with the plot size.

The heatmap_legend() function is used by passing it to the guide argument of the scale_fill_*() function.

Examples

The plot code used for the examples is below. The expand0() function is a convenience wrapper for expansion(mult = c(0,0)) to remove the space between the plot and the axes. theme(axis.line = element_blank()) removes the axis lines.

library(paletteer)

p <- ggplot(faithfuld, aes(waiting, eruptions, fill = density)) +
  geom_raster(interpolate = TRUE) +
  scale_x_continuous(expand = expand0()) +
  scale_y_continuous(expand = expand0()) +
  labs(x = "Time to Next Eruption (min)", y = "Eruption Time (min)",
       fill = "Density", title = "Old Faithful") +
  theme(axis.line = element_blank())

\ \ The ggplot default legend size:

p + scale_fill_paletteer_c("pals::ocean.tempo", direction = -1) + 
  theme_cr()

\ \ \ heatmap_legend() is passed to the guide argument of the scale_fill_*() function. It only has one argument, plot_scale, which corresponds to the plot_scale argument of the theme_cr function and has a default value of 1. The default creates a 6" by 4" plot with a font size of 12 pt.

p + scale_fill_paletteer_c("pals::ocean.tempo", direction = -1, 
                           guide = heatmap_legend()) + 
  theme_cr()

\ \ \ When the plot size is increased through the plot_scale argument in theme_cr(), the same value should be passed to heatmap_legend() to scale the color bar with the plot size.

The scaling is designed to be used with the crsave() function; here, it will appear as larger font and a larger color bar in the same plot size.

p + scale_fill_paletteer_c("pals::ocean.tempo", direction = -1, 
                           guide = heatmap_legend(plot_scale = 1.5)) + 
  theme_cr(plot_scale = 1.5)

\ \ \ To increase the size of the color bar without changing the size of the plot, just pass a value to plot_scale in heatmap_legend() without passing a value to theme_cr()

p + scale_fill_paletteer_c("pals::ocean.tempo", direction = -1, 
                           guide = heatmap_legend(plot_scale = 1.5)) + 
  theme_cr()

The appearance of legends in ggplot can be further adjusted with the override.aes argument of the guide_legend() function. For more information, see this article about the various applications of override.aes().

christyray/crthemes documentation built on April 14, 2023, 11:18 p.m.