Create a display of continuous versus categorical covariates

Get a single graphic of continuous covariate boxplots split by levels of different categorical covariates (cont_cat_panel()). Alternatively, get the component plots to be arranged by the user (cont_cat_panel_list())

Usage

cont_cat_panel(
  df,
  x,
  y,
  ncol = 2,
  tag_levels = NULL,
  byrow = FALSE,
  transpose = FALSE,
  points = NULL,
  ...
)

cont_cat_panel_list(df, x, y, transpose = FALSE, ...)

Arguments

df

a data frame to plot.

x

character col//title for the categorical covariates to plot on x-axis; see col_label().

y

character col//title for the continuous covariates to plot on y-axis; see col_label().

ncol

passed to pm_grid().

tag_levels

passed to patchwork::plot_annotation().

byrow

passed to pm_grid().

transpose

logical; if TRUE, output will be transposed to group plots by the categorical covariates rather than the continuous covariates.

points

passed to boxwork() when plotting categorical covariates.

...

additional arguments passed to cont_cat().

Value

cont_cat_panel() returns a list of plots arranged in graphics as a patchwork object using pm_grid(). cont_cat_panel_list() returns the same plots, but unarranged as a named list of lists.

When transpose is FALSE (default), plots in a single graphic are grouped by the continuous covariates (passed as y), and the names of the list reflect those names (e.g., WT). When transpose is TRUE, the graphics are grouped by the categorical covariates (passed as x) and the names of the list reflect those names (e.g. RF). See Examples.

Details

Pass ncol = NULL or another non-numeric value to bypass arranging plots coming from cont_cat_panel().

See also

eta_covariate(), eta_covariate_list()

Examples

data <- pmplots_data_id()
cont <- c("WT//Weight (kg)", "ALB//Albumin (mg/dL)", "AGE//Age (years)")
cats <- c("RF//Renal function", "CPc//Child-Pugh")

cont_cat_panel(data, x = cats, y = cont, tag_levels = "A")
#> $WT

#> 
#> $ALB

#> 
#> $AGE

#> 

cont_cat_panel(data, cats, cont)
#> $WT

#> 
#> $ALB

#> 
#> $AGE

#> 
cont_cat_panel(data, cats, cont, transpose = TRUE)
#> $RF

#> 
#> $CPc

#> 

l <- cont_cat_panel_list(data, cats, cont, transpose = TRUE)
names(l)
#> [1] "RF"  "CPc"
with(l$RF, WT/(ALB + AGE), tag_levels = "A")