ggbivariate(): Plot an outcome with several potential explanatory variables

library(GGally)
#> Loading required package: ggplot2
#> Registered S3 method overwritten by 'GGally':
#>   method from   
#>   +.gg   ggplot2

`GGally::ggbivariate()`

The purpose of this function is to easily plot a visualization of the bivariate relation between one outcome and several explanatory variables.

Basic example

Simply indicate the outcome and the explanatory variables. Both could be discrete or continuous.

data(tips)
ggbivariate(tips, outcome = "smoker", explanatory = c("day", "time", "sex", "tip"))

ggbivariate(tips, outcome = "total_bill", explanatory = c("day", "time", "sex", "tip"))

If no explanatory variables are provided, will take all available variables other than the outcome.

ggbivariate(tips, "smoker")

Customize plot title and legend title

ggbivariate(
  tips, "smoker", c("day", "time", "sex", "tip"),
  title = "Custom title"
) +
  labs(fill = "Smoker ?")

Customize fill colour scale

ggbivariate(tips, "smoker", c("day", "time", "sex", "tip")) +
  scale_fill_brewer(type = "qual")

Customize labels

ggbivariate(
  tips, "smoker", c("day", "time", "sex", "tip"),
  rowbar_args = list(
    colour = "white",
    size = 4,
    fontface = "bold",
    label_format = scales::label_percent(accurary = 1)
  )
)

Choose the sub-plot from which to get the legend

ggbivariate(tips, "smoker")

ggbivariate(tips, "smoker", legend = 3)

Change theme

ggbivariate(tips, "smoker") + theme_light()

Use mapping to indicate weights

d <- as.data.frame(Titanic)
ggbivariate(d, "Survived", mapping = aes(weight = Freq))

Use types to customize types of subplots

ggbivariate(
  tips,
  outcome = "smoker",
  explanatory = c("day", "time", "sex", "tip"),
  types = list(comboVertical = "autopoint")
)

For more customization options, you could directly use ggduo() (see also vig_ggally("ggduo")).

Joseph Larmarange

May 16, 2016