seastackplot

The goal of seastackplot is to create graphs that accurately represent the distribution and summary statistics of data.

Installation

You can install the development version of seastackplot from GitHub with:

# install.packages("remotes")
remotes::install_github("Al-Stu/seastackplot")

Creating a sea stack plot

The easiest way to plot a sea stack plot is using the sea_stack_plot() function. As an example, we will create a graph of how price differs between different diamond cuts using the ‘diamonds’ data set from the ggplot2 package. To create a sea stack plot using the sea_stack_plot(), you need to make sure your data is long format (see here for an explanation of long vs wide format data http://www.cookbook-r.com/Manipulating_data/Converting_data_between_wide_and_long_format/).

To plot a default sea stack plot with mean and standard deviation, the sea_stack_plot() function required the following variables: * data your long format data frame or tibble * data.column the name of the column your values are in * group.column the name of the column your group names are in

library(seastackplot)
#> Loading required package: ggplot2

sea_stack_plot(data = ggplot2::diamonds,
               data.column = 'price',
               group.column = 'cut',
               bin.width = 1000, # bin.width defaults to 1, for this data that is
                                # too high resolution so we set it to 1000
               show.confidence.int = T
               )

The sea_stack_plot() function also includes the following variables for plot customisation: * data.label the desired title for the data axis, if left NULL (default), will be the value of data.column

• bin.width the bin width being used to plot

• mean.size the size of the diamond showing the mean, measured in mm

• median.size the size of the point showing the median, measured in mm

• orientation “vertical” or “horizontal”, whether or not the plot is being plotted vertically (default) or horizontally

• show.mean whether the mean should be plotted

• show.median whether the median should be plotted

• show.standard.dev whether the standard deviation rectangle should be plotted

• show.confidence.int whether the confidence interval should be plotted

• show.quantiles whether the confidence interval should be plotted

• ci.line.length the length of the ‘ticks’ showing the upper and lower bounds of the confidence interval, measured on the same scale as the histogram

• ci.line.length the length of the ‘ticks’ showing the upper and lower bounds of the confidence interval, measured on the same scale as the histogram

sea_stack_plot(data = ggplot2::diamonds,
               data.column = 'carat',
               data.label = 'Carat',
               group.column = 'cut',
               bin.width = 0.2,
               show.mean = F, 
               show.standard.dev = F, # as mean and standard deviation shows by default need to manually silence them
               show.median = T,
               # show.quantiles = T,
               # quant.line.length = 300,
               median.size = 3
               )

An important statistic when working with sea stack plots is the maximum bin size within any group (i.e. the width of the count axis)