How do I iterate over time for different IDs in R?

Question

I'm trying to model assets depreciating over time in R. I have a dataset df=tibble(expand_grid(id=1:2, time=1:10), assets=id * 5000) , and want to show the assets falling by 250 in each period for each id. I've tried

input_data %>%
 group_by(id) %>%
 mutate(assets = case_when(time > 1 ~ lag(assets) - 250, TRUE ~ assets))

but this doesn't iterate in the way I want. I've also tried to use a for loop, but I've only figured out how to do this for the first ID -

for (i in 2:max(input_data$time)) {
   input_data$assets[i] <- input_data$assets[i-1] - 250
}

I'd really like advice as to how to do the same thing the for loop does for each ID - the input dataset should have about 100 IDs so an easy way to loop over them, or to split the dataset and apply the for loop to each group, would help. I'd especially appreciate an answer using tidyverse functions, since I know them much better.

Thank you. I hope this was clear enough.

Answer 1

I think this is what you want. I created a sequence assets_drop to ensure this code also works when time is date.

input_data %>% 
  group_by(id) %>% 
  mutate(assets_drop = seq(0, length.out = n(), by = 250), 
         assets_new = assets - assets_drop)

The output is

# A tibble: 20 x 5
# Groups:   id [2]
      id  time assets assets_drop assets_new
   <int> <int>  <dbl>       <dbl>      <dbl>
 1     1     1   5000           0       5000
 2     1     2   5000         250       4750
 3     1     3   5000         500       4500
 4     1     4   5000         750       4250
 5     1     5   5000        1000       4000
 6     1     6   5000        1250       3750
 7     1     7   5000        1500       3500
 8     1     8   5000        1750       3250
 9     1     9   5000        2000       3000
10     1    10   5000        2250       2750
11     2     1  10000           0      10000
12     2     2  10000         250       9750
13     2     3  10000         500       9500
14     2     4  10000         750       9250
15     2     5  10000        1000       9000
16     2     6  10000        1250       8750
17     2     7  10000        1500       8500
18     2     8  10000        1750       8250
19     2     9  10000        2000       8000
20     2    10  10000        2250       7750

Answer 2

Similar approach to @tivd's, here I define depreciation per period, then calculate the sum of all the prior depreciations, and then redefine assets to be it's original value minus depreciation to that point.

df %>% 
  group_by(id) %>%
  mutate(deprec_period = 250,
         deprec = pmin(assets, cumsum(lag(deprec_period, default = 0))),
         assets = assets - deprec)