TIL20230124

Correlation coefficient in R

cor

cor_body <- pg %>% 
  drop_na() %>% 
  group_by(species) %>% 
  summarise(across(.cols = bill_length_mm:flipper_length_mm, ~cor(., body_mass_g)))
cor_body

## # A tibble: 3 x 4
##   species   bill_length_mm bill_depth_mm flipper_length_mm
##   <fct>              <dbl>         <dbl>             <dbl>
## 1 Adelie             0.544         0.580             0.465
## 2 Chinstrap          0.514         0.604             0.642
## 3 Gentoo             0.667         0.723             0.711

cor_max <- cor_body %>% 
  summarise(across(.cols = bill_length_mm:flipper_length_mm, ~max(.))) %>% 
  summarise(.[which.max(.)])
cor_max

## # A tibble: 1 x 1
##   bill_depth_mm
##           <dbl>
## 1         0.723

cor_body %>% 
  filter_all(any_vars(. %in% c(cor_max$bill_depth_mm)))

## # A tibble: 1 x 4
##   species bill_length_mm bill_depth_mm flipper_length_mm
##   <fct>            <dbl>         <dbl>             <dbl>
## 1 Gentoo           0.667         0.723             0.711

cor_body %>% 
  filter(across(.cols = everything(),
                .fns = ~(.x %in% c(cor_max$bill_depth_mm))))

## # A tibble: 0 x 4
## # ... with 4 variables: species <fct>, bill_length_mm <dbl>,
## #   bill_depth_mm <dbl>, flipper_length_mm <dbl>

※ filter(), We cannot directly use across() in filter() because we need an extra step to combine the results. …

• https://dplyr.tidyverse.org/articles/colwise.html#filter
• https://community.rstudio.com/t/using-filter-with-across-to-keep-all-rows-of-a-data-frame-that-include-a-missing-value-for-any-variable/68442

rowAny <- function(x) rowSums(x) > 0
cor_body %>% 
  filter(
    rowAny(
      across(.cols = everything(),
             .fns = ~(.x %in% c(cor_max$bill_depth_mm))))
  )

## # A tibble: 1 x 4
##   species bill_length_mm bill_depth_mm flipper_length_mm
##   <fct>            <dbl>         <dbl>             <dbl>
## 1 Gentoo           0.667         0.723             0.711

cor_body %>% 
  filter(if_any(.cols = 2:4, ~(. %in% c(cor_max$bill_depth_mm))))

## # A tibble: 1 x 4
##   species bill_length_mm bill_depth_mm flipper_length_mm
##   <fct>            <dbl>         <dbl>             <dbl>
## 1 Gentoo           0.667         0.723             0.711