add_dataframe numerb of events

[phahn57]

I will simulate my actual environment, patients arriving at an hospital, six different departements. I have the number of arrivals per 15 minutes.

How can I add the numbers per time into the dataframe? Trying for one trajectory like this doesn't work:

Is that possible, or must I encode e.g. 3 times at 15 and 2 at 30 with:
15--1
15--1
15--1
30--1
30--1

[Enchufa2]

You basically need a single row per arrival; so, yes, you need to encode each row multiple times.

Simple example. Some sample data:

library(dplyr)

df <- data.frame(
  time = hms::as_hms(seq(7*60*60, 8*60*60, by=15*60)),
  n = rpois(5, 2)
)
df
#>       time n
#> 1 07:00:00 1
#> 2 07:15:00 3
#> 3 07:30:00 2
#> 4 07:45:00 1
#> 5 08:00:00 4

The basic way would be to just pick each row that number of times, then you no longer need the n column:

df |> 
  slice(rep(1:n(), n)) |> 
  select(-n)
#>        time
#> 1  07:00:00
#> 2  07:15:00
#> 3  07:15:00
#> 4  07:15:00
#> 5  07:30:00
#> 6  07:30:00
#> 7  07:45:00
#> 8  08:00:00
#> 9  08:00:00
#> 10 08:00:00
#> 11 08:00:00

However, if I understand correctly, you observed those many arrivals for the previous 15 min, which doesn't mean they arrived together at that exact time. So what I would do is to simulate this n number of arrivals for each 15 min period:

df |> 
  rowwise() |> 
  mutate(t = list(sort(runif(n, 0, 15*60)))) |> 
  tidyr::unnest_longer(t) |> 
  mutate(time = hms::as_hms(time - 15*60 + t)) |> 
  select(-n, -t)
#> # A tibble: 11 × 1
#>    time           
#>    <time>         
#>  1 06:59:09.127463
#>  2 07:01:56.243079
#>  3 07:07:01.216664
#>  4 07:12:30.103934
#>  5 07:23:14.985368
#>  6 07:23:17.406660
#>  7 07:33:35.005284
#>  8 07:47:42.738091
#>  9 07:51:04.753963
#> 10 07:56:24.461982
#> 11 07:57:48.193608

which solves a lot of problems that you may encounter when dealing with simultaneous events.

[phahn57]

Thanks I will try both solutions.