rproportions

How to calculate proportion of zeroes in dataframe by group? Presence/absence proportions


I have data of fish stomach contents (prey items).

In my original df, each fish (with a unique FID) had multiple rows(observations) - one row per unique prey taxon found. For example, if fish #10 had both daphnia and goby in its stomach, there were two rows for that same fish (one row with # of daphnia in that fish's stomach and one row for # of goby in that same stomach); if the fish only had daphnia in their stomach then they had one row; and so on.

I have converted my data from long to wide format to have one observation per row (one unique fish per row).

I am trying to calculate the proportion of empty stomachs by month (when totalnumPrey == 0).

Reproducible data (shortened; complete data has 488 observations):

structure(list(id = c("1001_28", "1001_29", "1001_30", "1001_31", 
    "1001_32", "1001_33", "1001_34", "1001_35", "1023_3", "614_1", 
    "614_3", "616_1", "616_3", "616_4", "616_5", "616_6", "824_23", 
    "824_24", "824_25", "824_26", "824_28", "824_29", "824_30", "824_31", 
    "824_32", "824_33", "824_35"), CRN = c(1001L, 1001L, 1001L, 1001L, 
    1001L, 1001L, 1001L, 1001L, 1023L, 614L, 614L, 616L, 616L, 616L, 
    616L, 616L, 824L, 824L, 824L, 824L, 824L, 824L, 824L, 824L, 824L, 
    824L, 824L), FID = c(28L, 29L, 30L, 31L, 32L, 33L, 34L, 35L, 
    3L, 1L, 3L, 1L, 3L, 4L, 5L, 6L, 23L, 24L, 25L, 26L, 28L, 29L, 
    30L, 31L, 32L, 33L, 35L), ac = c(2L, 2L, 1L, 1L, 1L, 1L, 0L, 
    0L, 1L, 2L, 2L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
    2L, 2L, 2L, 1L), mm = c(200L, 159L, 117L, 120L, 108L, 103L, 92L, 
    97L, 104L, 301L, 163L, 85L, 271L, 290L, 330L, 294L, 270L, 260L, 
    266L, 197L, 195L, 185L, 160L, 157L, 178L, 166L, 149L), gr = c(95, 
    44, 15.1, 16.1, 11, 10, 6.9, 7.9, 10.9, 418, 62, 6.8, 311, 453, 
    593, 395, 283, 275, 261, 96, 90, 90, 56, 50, 57, 62, 45.5), catch = c(2L, 
    2L, 4L, 4L, 4L, 4L, 2L, 2L, 1L, 3L, 3L, 1L, 5L, 5L, 5L, 5L, 15L, 
    15L, 15L, 15L, 15L, 15L, 15L, 15L, 15L, 15L, 14L), Daphnia = c(0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Byths = c(0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    18L, 79L, 71L, 8L, 73L, 0L, 38L, 39L), Chiro.Pupae = c(0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 255L, 7L, 0L, 576L, 590L, 536L, 576L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Empty = c(0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Chiro.Larvae = c(0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 5L, 38L, 0L, 9L, 0L, 0L, 0L), Amphipod = c(0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 6L, 0L, 0L, 0L, 4L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Isopod = c(0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 5L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L), Chironomidae = c(0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L), Hemimysis = c(0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L), Copepoda = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), Sphaeriidae = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), Chiro.Adult = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 74L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L), Trichopteran = c(0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L), UID.Fish = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), Chydoridae = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    200L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L), Cyclopoid = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L), Fish.Eggs = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L), EggMass = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L), Dreissena = c(0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 1L, 0L, 0L, 0L, 1L
    ), Goby = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Eurycercidae = c(0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), Hirudinea = c(0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
    0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), totalnumPrey = c(0, 0, 0, 
    0, 1, 0, 0, 0, 200, 262, 81, 0, 576, 595, 536, 582, 0, 0, 0, 
    19, 84, 110, 9, 82, 0, 38, 40), MONTH = c(11L, 11L, 11L, 11L, 
    11L, 11L, 11L, 11L, 11L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 8L, 8L, 
    8L, 8L, 8L, 8L, 8L, 8L, 8L, 8L, 8L), DAY = c(4L, 4L, 4L, 4L, 
    4L, 4L, 4L, 4L, 6L, 23L, 23L, 23L, 23L, 23L, 23L, 23L, 18L, 18L, 
    18L, 18L, 18L, 18L, 18L, 18L, 18L, 18L, 18L), empty = c("Empty", 
    "Empty", "Empty", "Empty", "Not_empty", "Empty", "Empty", "Empty", 
    "Not_empty", "Not_empty", "Not_empty", "Empty", "Not_empty", 
    "Not_empty", "Not_empty", "Not_empty", "Empty", "Empty", "Empty", 
    "Not_empty", "Not_empty", "Not_empty", "Not_empty", "Not_empty", 
    "Empty", "Not_empty", "Not_empty")), row.names = c(NA, -27L), class = c("data.table", 
    "data.frame"))

I haven't been able to figure out a way to calculate proportion using counts instead of actual values (since I need to count the 0 values by group and not use the actual 0 value to calculate the proportion).

I have tried the following:

example %>%
      group_by(empty, MONTH) %>%
      summarise(totalnumPrey = n()) %>%
      mutate(prop = n / sum(n))

This gives the following error:

Error in `mutate()`:
! Problem while computing `prop = n/sum(n)`.
ℹ The error occurred in group 1: empty = "Empty".
Caused by error in `sum()`:
! invalid 'type' (closure) of argument

I also tried this:

transform(example,
               perc = ave(totalnumPrey,
                          empty,
                          FUN = prop.table))

but this doesn't give me what I need...

Also this:

example %>%
  group_by(MONTH) %>%
  summarise(n = n()) %>%
  mutate(freq = n / sum(n))

which gives me proportion by month, not what I need (i.e. for June it's doing 127/362 = 0.35)...

I have tried many other ways from examples I found in other SO posts but still can't get what I need.

Is there a way I can calculate the proportion of empty vs non-empty stomachs by month?

I also need to do this for each prey type/taxon. For example, proportion of individual fish that contain "Isopod" and so on for each unique taxon in my data. Presence/absence type of proportions. I mainly want to do this by month first, but will eventually use other groupings.

When I had the data in long format, I was able to calculate proportion of each prey item within one fish stomach by using:

transform(a,
          perc = ave(number,
                     id,
                     FUN = prop.table))

data not included here.. but 'number' here being the total count of each unique prey taxa/group per stomach/fish & 'id' unique identifier I created to distinguish between different fish (since there were multiple rows for same fish).

I am happy to clarify anything that is not clear or add additional data if needed. I have searched online and in SO for a few days but still can't figure this out. Thank you in advance.


Solution

  • I think this is what you need.

    What we need to do is to count the number of times the column empty is equal to "Empty" per each group - so we can do this using sum(empty=="Empty") and then divide by the number of rows in that group n().

    library(dplyr)
    dat %>% 
        group_by(MONTH) %>% 
        summarise(
            prop_empty = sum(empty=="Empty")/n(), 
            prop_not_empy = sum(empty != "Empty")/n()
    )
    
    # A tibble: 3 × 3
      MONTH prop_empty prop_not_empy
      <int>      <dbl>         <dbl>
    1     6      0.143         0.857
    2     8      0.364         0.636
    3    11      0.778         0.222