Overlapping density distributions for multiple variables

eda_mdens generates overlapping density distributions for multiple batches.

Usage

eda_mdens(
  ...,
  p = 1L,
  tukey = FALSE,
  base = exp(1),
  grey = 0.6,
  alpha = 0.4,
  legend = TRUE,
  title = NULL,
  show.par = TRUE,
  kernel = "gaussian",
  cols = NULL,
  outline = FALSE
)

Arguments

...: Numeric vectors representing individual batches, or a dataframe and its continuous and categorical variables. Can also contain arguments passed to stats::density() like bw, n, from, to.
p: Power transformation to apply to all values.
tukey: Boolean determining if a Tukey transformation should be adopted (FALSE adopts a Box-Cox transformation).
base: Base used with the log() function if p = 0.
grey: Grey level to apply to plot elements (0 to 1 with 1 = black).
alpha: Fill transparency (0 = transparent, 1 = opaque).
legend: Boolean determining if a legend should be added to the plot.
title: Plot title. Defaults to "NULL".
show.par: Boolean determining if parameters such as power transformation should be displayed.
kernel: The kernel to be used. Can be one of the kernels supported by stats::density ("gaussian", "epanechnikov", "rectangular", "triangular", "biweight", "cosine", "optcosine"), or "sliding" for a custom sliding-window rectangular kernel estimator.
cols: A vector of colors for the density fills or one of the R built-in color palettes (see hcl.pals()). If NULL, colors default to "Dark 3".
outline: Boolean determining if outline should be shown without transparency.

Value

Does not return a value.

Details

This function extends eda_dens to allow for the comparison of more than two batches. Input can be a list of numeric vectors, individual numeric vectors passed directly, or a dataframe in long format.

Examples


# Example using individual vectors as input. A
# bandwidth of 0.5 is used for the density plots
# Plots are ordered following input order
set.seed(123)
x <- rnorm(100, 0, 1)
y <- rnorm(100, 2, 1.5)
z <- rnorm(100, 4, 2)
w <- rnorm(100, 8, 3)
eda_mdens(x, y, z, w, bw = 0.7)



# Example using a dataframe as input
# Plots are ordered alphabetically
df <- data.frame( values = c(x, y, z, w),
                  group = rep(c("x","y","z","w"), each = 100))
eda_mdens(df, values, group, bw=0.7)


# To specify an order when using a dataframe,
# use factors
df$group <- factor(df$group, levels = c(c("x","y","z","w")) )
eda_mdens(df, values, group, bw=0.7)


# Colors can be passed as a vector of color names or as a predefined
# hcl.pals() palette name. The fill transparency can be controlled with the
# alpha argument.
eda_mdens(df, values, group, cols = "OrRd", alpha = 0.6)


# Most density() arguments can be passed via ...
eda_mdens(df, values, group, bw = 0.5, n = 20, from=0, to=14,kernel="cosine")


# A built-in custom "sliding" window kernel is also available.
# Note that the bandwidth for "sliding" is in the variable's units and not
# the fraction of observations as implemented by density()
eda_mdens(df, values, group, bw = 5, n = 20, from=0, to=14, kernel="sliding")