Skip to contents

Missing data pattern

Source: R/md_pattern.R
md_pattern.Rd

Obtain a plot of the pattern of missing data and/or return the pattern as a matrix.

Usage

md_pattern(data, color = c(grDevices::grey(0.1), grDevices::grey(0.7)),
  border = grDevices::grey(0.5), plot = TRUE, pattern = FALSE,
  print_xaxis = TRUE, ylab = "Number of observations per pattern",
  print_yaxis = TRUE, legend.position = "bottom", sort_columns = TRUE,
  ...)

Arguments

data

data frame

color

vector of length two, that specifies the colour used to indicate observed and missing values (in that order)

border

colour of the grid

plot

logical; should the missing data pattern be plotted? (default is TRUE)

pattern

logical; should the missing data pattern be returned as matrix? (default is FALSE)

print_xaxis, print_yaxis

logical; should the x-axis (below the plot) and y-axis (on the right) be printed?

ylab

y-axis label

legend.position

the default position of legends ("none", "left", "right", "bottom", "top", "inside")

sort_columns

logical; should the columns be sorted by number of missing values? (default is TRUE)

...

optional additional parameters, currently not used

Value

With plot = TRUE and pattern = FALSE (default) a ggplot object. With pattern = TRUE and plot = FALSE a matrix with one row per missing data pattern (1: observed, 0: missing), the column Npat (number of observations with this pattern), and the last row Nmis (number of missing values per variable, and in total). With both set to TRUE a list with elements plot and pattern; with both FALSE, NULL.

Note

This function requires the ggplot2 package to be installed.

See also

See the vignette Visualizing Incomplete Data for more examples.

Examples

op <- par(mar = c(3, 1, 1.5, 1.5), mgp = c(2, 0.6, 0))
md_pattern(wideDF)

par(op)