How can I create a Correlation Matrix with multiple, specific variables?

Question

I have a very large data set, but want to do a correlation matrix with only 10 of the variables in that data set (and rename the variable names to something else)

My current code is looking like this:

#Correlation Function
data.cor = cor(df)

#Correlation Coefficient
data.cor = cor(df, method = "pearson", use = "complete.obs")


#Running Correlation w/ P-Values
install.packages("Hmisc")
library("Hmisc")
data.rcorr = rcorr(as.matrix(df))
data.rcorr

#
data.coeff = data.rcorr$r
data.p = data.rcorr$P

However, 1) it doesn't work 2) i am aware it is just reference the full data set

How can I make it so that I am referencing only those 10 variables and be able to rename the variables?

Thank you beforehand!

Answer 1

If you have a dataframe, d, you can directly subset in numerous ways. For example, using a vector of variable names, as below:

target_vars = c("v1", "v2", "v7", "v8", "v12")
cor(d[,target_vars], method = "pearson", use="complete.obs")

Output:

             v1         v2          v7         v8         v12
v1   1.00000000 0.57761512 -0.22465288  0.6571388  0.06053092
v2   0.57761512 1.00000000  0.08178772  0.4641090  0.14373593
v7  -0.22465288 0.08178772  1.00000000  0.3035047 -0.46410860
v8   0.65713875 0.46410902  0.30350472  1.0000000 -0.48583862
v12  0.06053092 0.14373593 -0.46410860 -0.4858386  1.00000000

Input:

set.seed(123)
d = setNames(as.data.frame(lapply(1:20, \(x) rnorm(10))), paste0("v",1:20))