Take a look at the ggpackets project page!
Easily build components of ggplots without sacrificing the ease of ggplot’s layer parameters and construction syntax.
install.packages("ggpackets")
or install the development version
devtools::install_github("dgkf/ggpackets", build_vignettes = TRUE)
There are plenty of ways to help contribute:
File issues!
Found a bug? Think the syntax looks ugly? Don’t like the name? Tell
me! Issues are the best
way to start the conversation.
Write documentation!
More resources always helps. Found a function unintuitive? Example
code and improved function descriptors would be helpful. If you use
the package and would feel comfortable writing about a topic not yet
covered in a vignette, feel free to contribute a new vignette for
it.
Write Unit Tests!
There’s some pretty sophisticated manipulations going on under the
hood to make everything as clean as possible, because of that it’s
important to make sure everything stays working the way we expect it
to. Unit test contributions always welcome!
Contribute Code!
Last but not least, code contributors are welcome. Reach out and get
in touch if you’re passionate about the goal of the project.
Define common ggplot layer sets together into a singled object. Connect all your plots with a single plot component definition and debug one central location. Build beautiful templates and save them once, reuse them easily and without having to abandon the ggplot construction framework.
ggpk_box_and_scatter <- ggpacket() +
geom_point(position = position_jitter(width = 0.4), alpha = 0.02) +
geom_boxplot(outlier.shape = NA, fill = NA, color = 'black') +
geom_text(stat = 'summary', vjust = -1, fun.data = function(d) c(
y = quantile(d, 0.75, names = F) + 1.5 * IQR(d),
label = length(d)
)) +
theme_linedraw() +
scale_color_distiller(palette = "Set1")
Now we can use that template repeatedly with a much simpler ggplot call
ggplot(diamonds, aes(x = cut, y = price, color = carat)) +
ggpk_box_and_scatter() +
ggtitle('Diamond price distribution by cut')
In addition to simply wrapping multiple ggplot2
layers, ggpackets
can streamline a number of complicated plotting scenarios such as
passing arguments to sets of layers, setting default argument values
with scoped overrides, routing aesthetic mappings to be reused within
specific layers for other aesthetics and scoping data usage over a set
of layers.
ggpk_labelled_heatmap <- function(...) {
ggpacket(...) %+%
geom_tile(.id = 'tile', color = NA, ...) %+%
geom_text(.id = c("text", "text1"), color = "black", vjust = -0.3,
fontface = "bold", ...) %+%
geom_text(.id = c("text", "text2"),
aes(label = sprintf("(%.1f)", ..fill..)),
color = "black", vjust = 1.1, ...) %+%
theme_void()
}
In this function we make use of a number of these specialized behaviors.
.id
parameters are set to tag specific layers with an identifier,
which can be used to prefix arguments to route them to a subset of
the ggpacket
layers. Multiple IDs can be used, and arguments will
filter down into that layer if they match any of the provided IDs.ggpacket(...)
, which will pass them
on as default values to all ggpacket
layers...<aesthetic>..
syntax.%+%
instead of the commonly-used +
to add layers
together, which allows ggpackets
to accept non-standard arguments
before ggplot sends us warnings about them.ggplot(as.data.frame(OrchardSprays)) +
aes(x = rowpos, y = colpos, label = treatment, fill = decrease) +
ggpk_labelled_heatmap(text.color = "white", text2.alpha = 0.5) +
ggtitle('Honeybee population decline in repellent trial grid')