Accelerating ggplot2
ggforce is a package aimed at providing missing functionality to ggplot2
through the extension system introduced with ggplot2 v2.0.0. Broadly speaking
ggplot2 has been aimed primarily at explorative data visualization in order to
investigate the data at hand, and less at providing utilities for composing
custom plots a la D3.js. ggforce is mainly an attempt to
address these "shortcomming" (design choices might be a better description). The
goal is to provide a repository of geoms, stats, etc. that are as well
documented and implemented as the official ones found in ggplot2.
The inclusion of any geom, stat, position etc in ggforce is not necessarily a
recommendation of their use. ggplot2 has been succesfull in being opinionated
about what functionality should be available. This is good as it insulates the
user from making bad decisions when analyzing their data (to a certain degree),
but it also makes it difficult to develop novel visualizations using the
ggplot2 API. ggforce on the other hand positions itself closer to the
"anything goes - the user is responsible for the quality of the output". Be very
aware of this responsibility! Bad visualizations lie about, distorts, and
obscurify the data behind them, both to you and the ones you share your
visualizations with.
ggforce is available on CRAN and can be installed in the regular way:
install.packages('ggforce')It is very possible that the GitHub repository contains new experimental
features not released on CRAN yet. To get a taste of the future use devtools
to install the development version:
if (!require(devtools)) {
install.packages('devtools')
}
devtools::install_github('thomasp85/ggforce')ggforce is by design a collection of features with the only commonality being
their tie to the ggplot2 API. Because of this an overview of all features
would get too long for a README. The package does contain a
vignette
where every feature is described and justified with examples and plots. There
should be a plot in the README of a visualization package though, so without
further ado:
ggplot(iris, aes(Petal.Length, Petal.Width, colour = Species)) +
geom_point() +
facet_zoom(x = Species == "versicolor")
Pull and feature requests are very welcome. Obviously PR's will lead to faster implementations than feature requests. I would like to urge requests to include the following if possible:
If a PR is for a new feature, it should be self contained, possibly using already implemented functionality if applicable. All exported functions should be documented following the style from ggplot2 using roxygen2 comment. You can credit yourself with the implementation in the documentation. If the feature concerns a visualization appraoch invented by others, please link to the article describing the approach.
If a feature is wished, but skill, time or other is lacking to create a full PR, please file an issue. The feature request should provide a detailed description of the nature of the feature, with links to relevant litterature describing the visualization type, as well as possible use cases to guide in designing the use cases.
For an overview of already requested features, enhancements, and bug fixes please consult the issues list. I will try to keep the labels up to date so it should be easy to identify whether an issue is related to any of the above.
ggforce is by no means unique. The ggplot2 ecosystem is flourishing
following the release of v2.0.0. To help keep on top of all ggplot2 extension
packages a web page has been created where developers can submit their packages.
Please explore https://www.ggplot2-exts.org to get an overview of the current
state of ggplot2 extensions.