Pattern Stripe - Parameters and Examples • ggpattern

suppressPackageStartupMessages({
  library(ggplot2)
  library(ggpattern)
})

Introduction `pattern = 'stripe'`

The stripe pattern

Pattern Parameters

aesthetic	description	default	possible values
pattern_colour	Stroke colour	‘grey20’	colour
pattern_fill	Fill colour	‘grey80’	colour
pattern_angle	Rotation angle	30	angle in degrees
pattern_density	Approx. fraction of area the pattern fills	0.2	value in range [0, 1] (fraction)
pattern_spacing	Spacing between repetitions of pattern	0.05	value in range [0, 1] (snpc units)
pattern_xoffset	Shift pattern along x axis	0	value in range [0, 1] (snpc units)
pattern_yoffset	Shift pattern along y axis	0	value in range [0, 1] (snpc units)
pattern_alpha	Alpha	NA	value in range [0, 1] or NA
pattern_linetype	Stroke linetype	1	linetype
pattern_size	Stroke linewidth	1	linewidth

There are also a number of parameters for extra control of legend sizing and aspect ratio adjustments. See the ‘Pattern Parameters - Common’ for more information.

Data

Standard data for all the example plots

df <- data.frame(trt = c("a", "b", "c"), outcome = c(2.3, 1.9, 3.2))
df
#>   trt outcome
#> 1   a     2.3
#> 2   b     1.9
#> 3   c     3.2

Vanilla ggplot2

ggplot(df, aes(trt, outcome)) +
  geom_col(aes(fill=trt),colour='black') +
  theme_bw() +
  labs(title = "Plain ggplot2")

Use the `{ggpattern}` geom

Use ggpattern::geom_col_pattern() instead of ggplot2::geom_col().
Set pattern = 'stripe'

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(aes(fill=trt),colour='black', pattern = 'stripe') +
  theme_bw() +
  labs(title = "ggpattern")

Use `coord_fixed()`

Due to the underlying nature of grid graphics and how ggplot2 sets up a plot, it is essential to use coord_fixed() in order to achieve consistent looking patterns. Without a fixed coordinate system, then ggpattern has to make guesses on some sizes, and patterns may look askew e.g. 90 degree angles in patterns may no longer appear properly!

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(aes(fill=trt),colour='black', pattern = 'stripe') +
  theme_bw() +
  labs(title = "ggpattern + coord_fixed()") + 
  coord_fixed(ratio = 1/2)

Bigger Legend Keys

Default legends in ggplot2 are usually too small to show off an example of the pattern, so it will usually be necessary to increase the key size.

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(aes(fill=trt),colour='black', pattern = 'stripe') +
  theme_bw() +
  labs(title = "ggpattern + coord_fixed()") + 
  coord_fixed(ratio = 1/2) + 
  theme(legend.key.size = unit(1.5, 'cm'))

The Density Aesthetic

The aesthetic pattern_density roughly corresponds to the fraction of the filled area which should be covered by the pattern.

In the following plot the density of striping is increased to 50% of the fill area.

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(
    aes(fill=trt),
    colour          = 'black', 
    pattern         = 'stripe',
    pattern_density = 0.5
  ) +
  theme_bw() +
  labs(title = "Fixed density of 0.5 (50% of the fill area)") + 
  coord_fixed(ratio = 1/2) + 
  theme(legend.key.size = unit(1.5, 'cm'))

The Density Aesthetic as a Mapped Aesthetic

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(
    aes(fill = trt, pattern_density = trt),
    colour          = 'black', 
    pattern         = 'stripe'
  ) +
  theme_bw() +
  labs(title = "Aesthetic Mapping of 'trt' to Density") + 
  coord_fixed(ratio = 1/2) + 
  theme(legend.key.size = unit(1.5, 'cm'))

The Density Aesthetic as a Mapped Aesthetic with Manual Scale

scale_pattern_density_manual() can be used to manually control how the variable is mapped to the density.

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(
    aes(fill = trt, pattern_density = trt),
    colour          = 'black', 
    pattern         = 'stripe'
  ) +
  theme_bw() +
  labs(title = "Aesthetic Mapping of 'trt' to Density") + 
  coord_fixed(ratio = 1/2) + 
  theme(legend.key.size = unit(1.5, 'cm')) + 
  scale_pattern_density_manual(values = c(a = 0.1, b=0.3, c=0.5))

The Spacing Aesthetic as a Mapped Aesthetic

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(
    aes(fill = trt, pattern_spacing = trt),
    colour          = 'black', 
    pattern         = 'stripe'
  ) +
  theme_bw() +
  labs(title = "Aesthetic Mapping of 'trt' to Spacing") + 
  coord_fixed(ratio = 1/2) + 
  theme(legend.key.size = unit(1.5, 'cm'))

The Fill Aesthetic as a Mapped Aesthetic

ggplot(df, aes(trt, outcome)) +
  geom_col_pattern(
    aes(fill = trt, pattern_fill = trt),
    colour          = 'black', 
    pattern         = 'stripe'
  ) +
  theme_bw() +
  labs(title = "Aesthetic Mapping of 'trt' to Pattern Fill") + 
  coord_fixed(ratio = 1/2) + 
  scale_pattern_fill_viridis_d() + 
  theme(legend.key.size = unit(1.5, 'cm'))

Technical Bits

Stripes are implemented as long thin polygons which are clipped to the boundaries of the elements of the geom.
The clipping of the stripe polygons is done by creating the stripes as sf objects, performing an sf::st_intersection() with the element boundary, and then converting the clipped stripes back into a grid::polygonGrob().
If lines had been used for the stripes, then there would be visual artefacts where the stripe intersects the boundary because the lineend graphic parameter can only be square, round etc, and cannot conform to the actual shape of the element boundary.

Pattern Stripe - Parameters and Examples

Introduction pattern = 'stripe'