corrmat: Correlation plot

In stopsack/khsmisc: Miscellaneous functions for epidemiology research

Correlation plot

Description

Performs pairwise Pearson or Spearman correlations.

Usage

corrmat(
  data,
  ...,
  method = "pearson",
  use = "pairwise.complete.obs",
  reorder = TRUE,
  digits = 2,
  fontsize = 4,
  legendpos = c(0.15, 0.35),
  cutpoints = c(-1, 0, 1),
  colors = c((viridis::viridis_pal(option = "cividis"))(10)[1], "white",
    (viridis::viridis_pal(option = "cividis"))(10)[8])
)

Arguments

data	Required. Data frame. Only numeric variables will be used; variables of other types will silently be dropped. To include categorical variables, coerce factors to numeric first. See examples.
...	Optional. Variables to correlate. If not provided, all numeric variables will be used. Supports tidy evaluation; see examples.
method	Optional. Correlation method. Defaults to "pearson".
use	How cor handles missing values. Defaults to a restriction to complete observations with all variables. See parameter use of cor for alternatives.
reorder	Perform hierarchical clustering in the correlation matrix? This will order variables by their correlation patterns with other variables. If turned off to FALSE, the order of variables in the dataset will be retained. Defaults to TRUE.
digits	Decimal digits for displayed correlation coefficients. Defaults to 2.
fontsize	Size for text in boxes and to the left of the boxes ("y axis"). Defaults to 4.
legendpos	(x, y) coordinates of color legend. Use legendpos = "none" to turn off the legend. Defaults to c(0.15, 0.35).
cutpoints	Correlation coefficient values that have a distinct color. Defaults to c(-1, 0, 1).
colors	Colors for the cutpoints. Defaults to blue (for negative), white (no correlation), and yellow (positive correlation) on the cividis color scale.

Value

ggplot. Can be modified with the usual ggplot commands, such as theme.

See Also

http://www.sthda.com/english/wiki/ggplot2-quick-correlation-matrix-heatmap-r-software-and-data-visualization

Examples

data(mtcars)

mtcars %>%
  corrmat(mpg, cyl, hp, wt, qsec)

# Can use tidy evaluation to select variables:
mtcars %>%
  corrmat(contains("a"), starts_with("c"))

# If "cyl" was a character, it would be excluded:
mtcars %>%
  dplyr::mutate(cyl_chr = as.character(cyl)) %>%
  corrmat(mpg, cyl_chr, hp, wt, qsec)

# To retain the character variable "cyl",
# convert to factor and then make numeric:
mtcars %>%
  dplyr::mutate(cyl_chr = as.character(cyl),
                cyl_chr = as.numeric(factor(cyl_chr))) %>%
  corrmat(mpg, cyl_chr, hp, wt, qsec)

stopsack/khsmisc documentation built on Sept. 22, 2023, 12:26 p.m.