marginal.boxplot: marginal.boxplot
In eheinzen/Ethan: A bunch of functions written by Ethan Heinzen

Description Usage Arguments Details Value See Also Examples

Plot boxplots on the margins of a scatterplot in ggplot2. This code is not totally written by Ethan; he adapted it from http://www.r-bloggers.com/scatterplot-with-marginal-boxplots/.

1 2	marginal.boxplot(data, x, y, by, xlim, ylim, xlab = "x", ylab = "y", bylab = "", width = 0.5, counts = FALSE, countsize = 5)

`data`	The data.frame to feed into ggplot2
`x`	A character string of the x-values to use in 'data'. Will be coerced to numeric if it isn't already.
`y`	A character string of the y-values to use in 'data'. Will be coerced to numeric if it isn't already.
`by`	A character string of the factor to use in 'data' for color/fill. If missing, defaults to no grouping. Will be coerced to factor if it isn't already.
`xlim`	The x-limits. If missing, defaults to a good view.
`ylim`	The y-limits. If missing, defaults to a good view.
`xlab`	The x-label. Defaults to 'x'. See details.
`ylab`	The y-label. Defaults to 'y'. See details.
`bylab`	The label for the by-variable. Defaults to "".
`width`	The width of jitter. Default 0.1. A warning is issued if > 0.3.
`counts`	Display the counts of each 'by' group in the top right corner? If 'by' is missing, displays nrow(data).
`countsize`	If the counts are displayed, what size should they be? Default = 5 - floor(# levels / 4).

I considered using arguments that don't require character strings, but I think this way might just be safer. The only downside is that I need to fill the xlab and ylab so they're not so dang ugly. That's why the defaults are to 'x' and 'y'. The plus-side is that I have fewer if-else's when specifying '+xlab()+ylab()' to 'ggplot()'. It's also easier (in my opinion) to catch when the variable names aren't in 'data'.

Nothing is returned; however, the ggplot objects are printed.

http://www.r-bloggers.com/scatterplot-with-marginal-boxplots/

1
2
3

marginal.boxplot(pizza.data, x = "height", y = "time_to_fu")
marginal.boxplot(pizza.data, x = "height", y = "time_to_fu", by = "sex",
 xlab = "Height", ylab = "Time to Follow-Up", bylab = "Sex", counts = TRUE)