ggsvg - Use SVG as points in ggplot

ggsvg is an extension to ggplot to use SVG images for points.

Variables may be aesthetically mapped to features within the SVG using CSS selectors via the css() helper function.

What’s in the box

• geom_point_svg() for plotting points with SVG as the glyph (This is a direct analogue to geom_point())
• scale_svg_*() functions for controlling the aesthetic mapping.
  • scale_svg_default() is a sensible default for most plots.
  • scale_svg_* are a shadow set of ggplot2::scale_*() functions with adaptations needed for css() selectors as aesthetics.
  • E.g. scale_svg_fill_brewer() is a direct analogue for ggplot2::scale_fill_brewer()

Installation

Install from GitHub.

The {rsvg} package is used to convert SVG into an R raster object. This requires at least rsvg(>= 2.3.0).

# install.package('remotes')
install.packages('rsvg')
remotes::install_github('coolbutuseless/ggsvg')

Simple plot

svg_url <- 'https://www.svgrepo.com/download/289000/jellyfish.svg'
svg_txt <- paste(readLines(svg_url), collapse = "\n")

grid::grid.draw( svg_to_rasterGrob(svg_txt) )

test_df <- data.frame(
  x = runif(10), 
  y = runif(10), 
  count = sample(3:5, 10, T),
  type  = sample(c('a', 'b', 'c'), 10, T))

test_df
#>             x         y count type
#> 1  0.26550866 0.2059746     3    b
#> 2  0.37212390 0.1765568     5    b
#> 3  0.57285336 0.6870228     3    a
#> 4  0.90820779 0.3841037     3    c
#> 5  0.20168193 0.7698414     3    a
#> 6  0.89838968 0.4976992     3    c
#> 7  0.94467527 0.7176185     4    b
#> 8  0.66079779 0.9919061     3    b
#> 9  0.62911404 0.3800352     3    b
#> 10 0.06178627 0.7774452     4    b

ggplot(test_df) + 
  geom_point_svg(aes(x, y), svg = svg_txt) + 
  theme_bw()

Simple plot with mapped size aesthetic

ggplot(test_df) + 
  geom_point_svg(aes(x, y, size = type), svg = svg_txt) + 
  theme_bw()

Mapping Aesthetics to SVG features with CSS Selectors

Aesthetic values are mapped to SVG features with CSS Selectors.

Snowman SVG

Here is a simple SVG consisting of 2 stacked circles - a big circle on the bottom and a small circle resting on top.

snowman_txt <- '
  <svg viewBox="0 0 100 100 ">
    <circle id="top" cx="50" cy="20" r="20" fill="brown" stroke="black" />
    <circle id="bot" cx="50" cy="70" r="30" fill="brown" stroke="black" />
  </svg>
  '

grid::grid.draw( svg_to_rasterGrob(snowman_txt, width=800, height=800) )

css() helper function

Use the css() helper function to target aesthetics at selected elements within and SVG using css(selector, property = value)

E.g.

• css("rect.big", stroke = x)
  • Targets <rect> elements with class = "big"
  • Map values in x in data.frame to the SVG stroke property for these targetted elements.

Example

In the following example, two css() selectors are used within the geom_point_svg() call:

• css("circle#top", fill=type)
  • Targets <circle> elements with id = "top"
  • Map values in type in data.frame to the SVG fill property for these targetted elements.
• css("circle#bot", stroke='brown')
  • Targets <circle> elements with id = "bot"
  • Set a constant value of brown for the SVG stroke property for these targetted elements.
• css("circle", 'stroke-width'=10)
  • Targets <circle> elements
  • Set a constant value of 5 for the SVG stroke-wdith property for these targetted elements.

To configure how the variable is mapped to the property on the selected target, you can either use:

• scale_svg_default() for reasonable defaults
• scale_svg_*() family of functions
  • Note: the aesthetic argument must match exactly the css(...) call used in the geom_point_svg() call.

snowman_txt <- '
  <svg viewBox="0 0 100 100 ">
    <circle id="top" cx="50" cy="20" r="20" fill="brown" stroke="black" />
    <circle id="bot" cx="50" cy="70" r="30" fill="brown" stroke="black" />
  </svg>
  '


ggplot(test_df) + 
  geom_point_svg(
    aes(x, y, css("circle#top", fill = type)),
    css("circle#bot", stroke = 'brown'),
    css("circle", 'stroke-width'=10),
    svg = snowman_txt
  ) +
  theme_bw() + 
  scale_svg_fill_brewer(aesthetics = css("circle#top", fill = type), palette = 'Dark2')

Acknowledgements

• R Core for developing and maintaining the language.
• CRAN maintainers, for patiently shepherding packages onto CRAN and maintaining the repository