Calculating non-continuous duration period in R-CodePudding

I have a large dataset, with numerous behavioral observations and I am trying to calculate the duration of each individual behaviour daily.

Here's a sample of it for 2022-06-04

library(tidyverse)
data_tibble <- tibble::tribble(
  ~bird_ID,       ~longitude,        ~latitude,        ~date, ~Position_Burst_ID,            ~timestamp,  ~longitude_mean,   ~latitude_mean,            ~Dist,
  "048T", 58.4836921691895, 68.8644866943359, "2022-06-04",               5121, "2022-06-04 09:49:38", 58.4966227901897, 68.8623445334237, 571.205432837044,
  "048T", 58.4836730957031, 68.8645324707031, "2022-06-04",               5122, "2022-06-04 10:19:43", 58.4966227901897, 68.8623445334237, 574.044222782275,
  "048T", 58.4962844848633, 68.8621215820312, "2022-06-04",               5123, "2022-06-04 10:49:36", 58.4966227901897, 68.8623445334237, 28.2914837022366,
  "048T", 58.4961318969727, 68.8620452880859, "2022-06-04",               5124, "2022-06-04 11:19:44", 58.4966227901897, 68.8623445334237, 38.7040772369604,
  "048T", 58.4965591430664,  68.862174987793, "2022-06-04",               5125, "2022-06-04 11:49:37", 58.4966227901897, 68.8623445334237, 19.0458862694959,
  "048T", 58.4965972900391, 68.8622360229492, "2022-06-04",               5126, "2022-06-04 12:19:48", 58.4966227901897, 68.8623445334237, 12.1226274623415,
  "048T", 58.4966316223145,  68.862419128418, "2022-06-04",               5127, "2022-06-04 12:49:39", 58.4966227901897, 68.8623445334237, 8.31144228341025,
  "048T", 58.4971313476562, 68.8621292114258, "2022-06-04",               5128, "2022-06-04 13:19:39", 58.4966227901897, 68.8623445334237,  31.485137709381,
  "048T",  58.494312286377, 68.8651657104492, "2022-06-04",               5129, "2022-06-04 13:49:41", 58.4966227901897, 68.8623445334237,  327.46020021598
)

^{Created on 2023-01-10 with}

Any idea on how to approach this?

This is the output I get for 2022-06-04:

outpout
#> # A tibble: 1 × 2
#>   date       duration      
#>   <chr>      <drtn>        
#> 1 2022-06-04 23.00056 hours

But I want to calculate the duration of the low points in the graph, so it should be closer to 9hrs, not 23hrs.

Edit1: I forgot to mention that I want to calculate that on a daily basis.

Edit2: I added a semi desirable output.

CodePudding user response：

Here's my approach with base function and for loop

table_recap <- data_tibble[1,]
table_recap$diff_time <- NA
tab2 <- table_recap

for (i in unique(data_tibble$bird_ID)){
  
  subset_data_tibble <- data_tibble[data_tibble$bird_ID==i,]
  
  subset_data_tibble <- subset_data_tibble[order(subset_data_tibble$timestamp),]
  
  time1 <- subset_data_tibble[1,]
  time2 <- subset_data_tibble[nrow(subset_data_tibble),]
  
  table_recap$diff_time <- difftime(time2$timestamp, time1$timestamp)
  
  tab2 <- rbind(tab2, table_recap)
  
}

For each "bird_ID", results will be stored in the table tab2

CodePudding user response：

Ideally, we should have sample data with multiple dates and not only for 2022-06-04

library(tidyverse)

data_tibble %>% 
  type_convert() %>%  
  group_by(date) %>% 
  summarise(
    duration = sum(if_else(Dist <= 30, lead(timestamp) - timestamp, 0))
  ) 

# A tibble: 1 × 2
  date       duration     
  <date>     <drtn>       
1 2022-06-04 120.1667 mins