correlation between columns by group

0

votes

How do I calculate correlations between one column and all other columns in a data frame in R without using column names? I tried to use ddply and it works if I use just two column names i.e.

library(plyr)
ddply(iris, ~Species, summarize, cormat=cor(Sepal.Length,Petal.Width))

But how to get correlations of column 1 with all other columns, broken down by Species, without using the column names?

r correlation

4

votes

Maybe like this? It produces a correlation matrix for each species.

by(iris[,1:4], iris$Species, cor)

1

votes

You can do this with dplyr using

library(dplyr)
cormat_res <- iris %>%
   group_by(Species) %>%
   do(cormat = cor(select(., -matches("Species"))))


> cormat_res[[2]]
[[1]]
             Sepal.Length Sepal.Width Petal.Length Petal.Width
Sepal.Length    1.0000000   0.7425467    0.2671758   0.2780984
Sepal.Width     0.7425467   1.0000000    0.1777000   0.2327520
Petal.Length    0.2671758   0.1777000    1.0000000   0.3316300
Petal.Width     0.2780984   0.2327520    0.3316300   1.0000000

[[2]]
             Sepal.Length Sepal.Width Petal.Length Petal.Width
Sepal.Length    1.0000000   0.5259107    0.7540490   0.5464611
Sepal.Width     0.5259107   1.0000000    0.5605221   0.6639987
Petal.Length    0.7540490   0.5605221    1.0000000   0.7866681
Petal.Width     0.5464611   0.6639987    0.7866681   1.0000000

[[3]]
             Sepal.Length Sepal.Width Petal.Length Petal.Width
Sepal.Length    1.0000000   0.4572278    0.8642247   0.2811077
Sepal.Width     0.4572278   1.0000000    0.4010446   0.5377280
Petal.Length    0.8642247   0.4010446    1.0000000   0.3221082
Petal.Width     0.2811077   0.5377280    0.3221082   1.0000000

1

votes

As of

packageVersion("dplyr")
[1] ‘1.0.2’

The result of the code suggested in one of the answers returns a tibble

iris %>%
     group_by(Species) %>%
     do(cormat = cor(select(., -matches("Species"))))
# A tibble: 3 x 2
# Rowwise: 
  Species    cormat           
  <fct>      <list>           
1 setosa     <dbl[,4] [4 × 4]>
2 versicolor <dbl[,4] [4 × 4]>
3 virginica  <dbl[,4] [4 × 4]>

To get the data into a rectangular shape, you can

iris_cor <- iris %>%
     group_by(Species) %>%
     do(cormat = cor(select(., -matches("Species")))) %>%
     pull(cormat) %>% melt

You will have the levels of Species codified on L1 variable.

           Var1         Var2     value L1
1  Sepal.Length Sepal.Length 1.0000000  1
2   Sepal.Width Sepal.Length 0.7425467  1
3  Petal.Length Sepal.Length 0.2671758  1
4   Petal.Width Sepal.Length 0.2780984  1
...

I am sure there's a cleaner way of doing this with unnest() and its friends, but couldn't figure out yet. Hoping this gets noticed and posts a better solution

correlation between columns by group

3 Answers