Without getting into the details of the application, I need to run the FSA::depletion(function) in R on a dataframe where the function needs to run iteratively to find a ‘depletion’ value for X number of rows and then proceed to run the operation on the next X number of rows. A dataframe will look something like this:
df <- data.frame(site = c('site1','site1','site1','site2','site2','site2','site3','site3','site3'),
catch = c(30,10,3,40,12,3,10,3,1),
effort = c(600,550,500,800,800,700,400,425,375))
Normally, I would run something like this to find the depletion value on the each site group:
library(dplyr)
library(FSA)
df %>%
group_by(site) %>%
depletion(catch, effort, method = 'Leslie')
However, it seems as though this function simply won’t work with dplyr as it won’t recognize ‘catch’ or ‘effort’ as vectors from df. The function seems to only be able to run when using the dataframe$vector notation such as the following:
depletion(df$catch, df$effort, method = 'Leslie')
Would anyone be able to suggest a way to make this work or a workaround? My next thought would be to add a count column and run the whole thing as a for loop but that sounds like a bad process to make this work.
You could do something like this:
library(purrr)
res <- df %>%
nest_by(site) %>%
mutate(d = list(depletion(data$catch, data$effort, method = 'Leslie')))
You still have to decide how to unpack the results/which bits you want to save: res$d is a list, if you look at res$d[[1]] you can see that it’s a complicated object with a lot of components.
You can use the broom package to tidy the output into your data frame:
library(FSA)
library(dplyr)
library(broom)
df %>%
nest_by(site) %>%
mutate(d = list(depletion(data$catch, data$effort, method = 'Leslie')),
No=list(summary(d, parm="No")),
q=list(summary(d, parm="q"))) %>%
unnest(col=c(No, q))
Gives:
# A tibble: 3 × 5
# Groups: site [3]
site data d No[,1] [,2] q[,1] [,2]
<chr> <list<tibble[,2]>> <list> <dbl> <dbl> <dbl> <dbl>
1 site1 [3 × 2] <depletin> 46 0.818 0.00109 0.0000315
2 site2 [3 × 2] <depletin> 57.0 0.131 0.000878 0.00000334
3 site3 [3 × 2] <depletin> 14.3 0.308 0.00174 0.0000617
Where the second columns of No and q are the standard errors, and
df %>%
nest_by(site) %>%
mutate(d = list(depletion(data$catch, data$effort, method = 'Leslie')),
lm = list(tidy(d$lm))) %>%
unnest(col=lm)
Gives
# A tibble: 6 × 8
# Groups: site [3]
site data d term estimate std.error statistic p.value
<chr> <list<tibble[,2]>> <list> <chr> <dbl> <dbl> <dbl> <dbl>
1 site1 [3 × 2] <depletin> (Intercept) 0.0502 0.000909 55.2 0.0115
2 site1 [3 × 2] <depletin> K -0.00109 0.0000315 -34.6 0.0184
3 site2 [3 × 2] <depletin> (Intercept) 0.0500 0.000126 396. 0.00161
4 site2 [3 × 2] <depletin> K -0.000878 0.00000334 -263. 0.00242
5 site3 [3 × 2] <depletin> (Intercept) 0.0249 0.000584 42.6 0.0149
6 site3 [3 × 2] <depletin> K -0.00174 0.0000617 -28.2 0.0226