2

I am trying to figure out the most efficient way to achieve a series of goals to group my data, summarize columns, and mutate a new column based on the summary.

With the example data below, I want to:

  1. mutate a new column "sum", which would be the sum of "count" , group_by(site, trmt, id, species)
  2. calculate the relative abundance of each species, group_by(id).

This post comes close to helping me out, but I'm not trying to summarise(across()) multiple columns: dplyr: group_by, sum various columns, and apply a function based on grouped row sums?

How would you work through this with pipes in dplyr to get from ''df_have'' to ''df_want''?

Thank you!

site <- c("X", "Y", "Y", "X", "X", "X", "Y", "X", "Y", "X", "Y", "Y", "X", "X", "X", "Y", "X", "Y")
trmt <- c("yes", "yes", "no", "no", "yes", "no", "no", "yes", "yes", "yes", "yes", "no", "no", "yes", "no", "no", "yes", "yes")
id <- c(1, 2, 3, 4, 5, 6, 7, 8, 9, 1, 2, 3, 4, 5, 6, 7, 8, 9)
species <- c("a", "b", "a", "c", "d", "a", "e", "b", "d", "a", "b", "m", "c", "p", "a", "q", "r", "d")
count <- c(28, 17, 7, 8, 2, 9, 1, 5, 3, 12, 4, 18, 3, 30, 12, 21, 18, 6)
extra <- c("A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B")


df_have <- cbind(site, trmt, id, species, count, extra) 
df_have <- as.data.frame(df_have)
df_have


site1 <- c("X", "Y", "Y", "X", "X", "Y", "Y",  "X", "X", "Y" )
trmt1 <- c("yes", "yes", "no", "yes", "no", "no", "no", "yes", "yes", "yes" )
id1 <- c(1, 2, 3, 3, 4, 5, 5, 6, 7, 7, 8, 8, 9)
species1 <- c("a", "b", "a", "m", "c", "d", "p", "a", "e", "q", "b", "r", "d" )
sum <- c(40, 21, 7, 18, 11, 2, 30, 21, 1, 21, 5, 18, 9)
relabund <- c(100, 100, 38.9, 61.1, 100, 6.25, 93.75, 100, 4.54, 95.45, 27.74, 78.26, 100)

df_want <- cbind(site1, trmt1, id1, species1, sum, relabund) 
df_want <- as.data.frame(df_want)
df_want
slane
  • 61
  • 5

1 Answers1

3

Here is a dplyr option

library(dplyr)
df_have %>%
    group_by(site, trmt, id, species) %>%
    summarise(sum = sum(as.integer(count)), .groups = "drop") %>%
    group_by(id) %>%
    mutate(relabund = sum / sum(sum) * 100) %>%
    ungroup() %>%
    arrange(id, species)
## A tibble: 13 x 6
#   site  trmt  id    species   sum relabund
#   <chr> <chr> <chr> <chr>   <int>    <dbl>
# 1 X     yes   1     a          40   100   
# 2 Y     yes   2     b          21   100   
# 3 Y     no    3     a           7    28   
# 4 Y     no    3     m          18    72   
# 5 X     no    4     c          11   100   
# 6 X     yes   5     d           2     6.25
# 7 X     yes   5     p          30    93.8 
# 8 X     no    6     a          21   100   
# 9 Y     no    7     e           1     4.55
#10 Y     no    7     q          21    95.5 
#11 X     yes   8     b           5    21.7 
#12 X     yes   8     r          18    78.3 
#13 Y     yes   9     d           9   100

The last arrange() command is just to match your expected output; if the order doesn't matter you can omit. Also note that the data in the count column are characters, so we need to convert to integer first; this is something that should probably be fixed upstream.

Maurits Evers
  • 49,617
  • 4
  • 47
  • 68