cor_list: Correlation List

In jashu/corxplor: Correlation Exploration

Description

Return the correlations between the columns of a matrix or data frame as a "cor_list" object.

Usage

cor_list(x, y = NULL, use = "pairwise", method = "pearson")

Arguments

x	a numeric matrix or data frame.
y	NULL (default) or a matrix or data frame with compatible dimensions to x. The default is equivalent to y = x (but more efficient).
use	an optional character string giving a method for handling missing values. This must be (an abbreviation of) one of the strings "everything", "all.obs", "complete.obs", "na.or.complete", or "pairwise.complete.obs" (default). See cor for explanation of these options.
method	a character string indicating which correlation coefficient is to be computed. One of "pearson" (default), "kendall", or "spearman": can be abbreviated.

Details

The cor function from the R stats package returns bivariate correlations as a matrix of values, which can be difficult to parse when there are many correlations. cor_list converts the correlation matrix into a list of all row-column pairs, removing the diagonal entries, and stores the type of correlation coefficient (Pearson's r, Kendall's tau, or Spearman's rho) as an attribute in a cor_list object with specialized print and summarize methods. The print method automatically outputs unique bivariate relationships, sorted from strongest to weakest. The summarise.cor_list method enables the use of select_helpers expressions to return specific bivariate relationships that you want to look at.

cor_list differs from R's native cor in the following ways:

1. x (and, if given, y) must be a matrix or data frame with named columns; unnamed vectors are not allowed. (You can still pass data for a single variable as x or y, but it must be in the form of a one-column matrix or data frame rather than a numeric vector.)

2. The default argument for use is pairwise-complete observations instead of "everything".

3. The cor_list object stores the type of correlation coefficient as an attribute.

4. The cor_list object has special summarize methods for selectively viewing relationships of interest. See summarize.cor_list.

Value

A cor_list object with the following three vectors:

x

variables that form the rows of the correlation matrix

y

variables that form the columns of the correlation matrix

coef

the correlation coefficient between corresponding elements of x and y

and the attribute "coef" indicating the statistic that is being returned (Pearson's r, Kendall's tau, or Spearman's rho).

See Also

cor_boot, summarise.cor_list

Examples

# Create a correlation list for the numeric variables from the iris data set
iris_cors <- cor_list(iris[,-5])

# Print returns all unique bivariate relationships sorted by strength
iris_cors

# Look at all correlations with Sepal.Length
summarize(iris_cors, x = Sepal.Length)

# Look at all correlations between sepal measurements and petal measurements
summarize(iris_cors, x = starts_with("Sepal"), y = starts_with("Petal"))

# Look at all correlations with Sepal.Length, excluding Sepal.Width
summarize(iris_cors, x = Sepal.Length, y = -Sepal.Width)

# Look at all correlations in their original order
summarize(iris_cors, sort = FALSE)

jashu/corxplor documentation built on Dec. 10, 2019, 7:09 p.m.