splot

An R package to ease data visualization.

The aim of this package is to make visualization an early part of the data analysis process by automating a few common plotting tasks.

In terms of design, it has three general principles:

• Flexibility: splot prefers to try to make a reasonable plot from any input, rather than erroring out.
• Minimal specification: You should be able to make most plots with just a formula input, and the same formula should generally be compatible with multiple plot types, regardless of variable types.
• Tweakability: Though splot is focused on quick, automated plotting, you should be able to adjust any aspect of the display with additional arguments, if you find a plot you want to display elsewhere.

features

By entering a formula as the first argument in the splot function (e.g., splot(y ~ x)), you can make

• Density distributions (overlaid on histograms when there is no by variable)
• Scatter plots with prediction lines
• Bar or line graphs with error bars

For each type, multiple y variables or data at levels of a by variable are shown in the same plot frame,
and data at levels of one or two between variables are shown in separate plot frames, organized in a grid.

resources

• Introduction
• Documentation
  
• Style Guide
  
• Gallery
• Applied examples: Exploring Data | Refining a Result

installation

Download R from r-project.org.

Release (version 0.5.2)

install.packages('splot')

Development (version 0.5.3)

install.packages('remotes')
remotes::install_github('miserman/splot')

Then load the package:

library(splot)

examples

Make some data: random group and x variables, and a y variable related to x:

group = rep(c('group 1', 'group 2'), 50)
x = rnorm(100)
y = x * .5 + rnorm(100)

The distribution of y:

splot(y)

A scatter plot between y and x:

splot(y ~ x)

Same data with a quadratic model:

splot(y ~ x + x^2 + x^3)

Same data separated by group:

splot(y ~ x * group)

Could also separate by median or standard deviations of x:

splot(y ~ x * x)
splot(y ~ x * x, split='sd')

Summarize with a bar plot:

splot(y ~ x * group, type='bar')

Two-level y variable with a probability prediction line:

# make some new data for this example:
# a discrete y variable and related x variable:
y_bin = rep(c(1, 5), 50)
x_con = y_bin * .4 + rnorm(100)

# lines = 'prob' for a prediction line from a logistic model:
splot(y_bin ~ x_con, lines = 'prob')