In wcmbishop/gogoplot: Interactive ggplot2 gadgets

gogoplot

gogoplot provides a simple GUI for building a ggplot data visualization (using a shiny HTML gadget). A data viz GUI is nothing new thing. What gogoplot aims to do is to dynamically build and return the code used to make a visualization, which makes it reproducible. This is also meant as an education tool to help introduce folks to some of the core functionality, syntax, and concepts of ggplot2.

This package is a work in progress. I would love help improving it! Check out the "Contribute" section if you're interested.

Install

gogoplot is only currently only available on github, as a development version. You can install this with the code below, using the devtools package.

# install.packages("devtools")
devtools::install_github("wcmbishop/gogoplot", force = TRUE)

Example

The current version of gogoplot has just one function - gogoplot(). When you pass in a data-frame, a gadget UI will appear (either in your RStudio Viewer pane or in a pop-up window). From there, you can interactively build up a visualization. At any point you can show the ggplot2 code that is being generated, and when you're finished you can return that code. If you're using RStudio, the generated plot code is automatically inserted wherever your cursor is located.

library(ggplot2)
library(gogoplot)
gogoplot(mtcars)

RStudio Addin

The package also installs a handy RStudio IDE add-in, available in the "Addins" drop-down. To use this, highlight a data-frame object in your code and select the gogoplot add-in to run the gadget.

Contribute

This is the very beginning of this package, and I would love help and feedback in developing it. The core concept is to provide simple GUIs for common R activities, which can then return the generated code as a result. As mentioned above, this seems useful both as a learning tool and a way to generate reproducible code. I'd ideally like to build these GoGo gadgets for some of the most valuable and frequently used tidyverse tools/workflows below. I'm starting with ggplot2 because it's what got me into R.

• plotting with ggplot2
• data manipulation with dplyr
• data tidying with tidyr

The core piece to making these gadgets robust and scalable is to implement the programmatic code generation well. This has recently been improved to capture both plot updates and plot code in one step, with a new %++% operator. This function adds a plot layer and captures the code in an object attribute (as a character vector). This relies heavily on the amazing rlang package and quasiquotation under the hood.

If you would like to help, please post a new github issue or send me a pull-request.

wcmbishop/gogoplot documentation built on May 29, 2019, 10:39 a.m.