Camera data wrangling

The use of camera traps in ecological studies has become increasingly popular for monitoring wildlife. Managing and analyzing camera trap data efficiently is important for extracting meaningful and accurate insights. In this vignette, we will explore how to perform camera data wrangling, specifically focusing on the ABMI’s Ecosystem Health 2014 data set.

Use the wt_download_report() function to retrieve the main report for the CAM sensor in the Ecosystem Health 2014 project:

eh14_raw <- wt_download_report(
  project_id = 205, 
  sensor_id = "CAM",
  report = "main", 
  weather_cols = FALSE
)

Evaluate independent detections with wt_ind_detect():

# Back to the Ecosystem Health 2014 data.

eh14_detections <- wt_ind_detect(
  x = eh14_raw, 
  threshold = 30,
  units = "minutes",
  remove_human = TRUE, 
  remove_domestic = TRUE 
)

glimpse(eh14_detections, width = 75) 

313 independent detections in this data set, when using a threshold of 30 minutes.

The output from wt_ind_detect() gave us some useful information. But we probably need to do additional wrangling for our data to be in the proper format for certain modeling techniques (e.g. habitat modeling, occupancy). For example, we want to evaluate the number of detections in a specified time interval (e.g. daily, weekly, or monthly), including zeroes.

Summarise your camera data

With wt_summarise_cam() you can get:

• The output from wt_ind_detect() (e.g. the object eh14_detections)
• Your raw data (e.g. the object eh14_raw)
• The time interval you’re interested in (e.g. weekly)
• The variable you’re interested in (e.g. detections, presence/absence)
• The desired output format (‘wide’ or ‘long’)

# A call to `wt_summarise_cam()`:

eh14_summarised <- wt_summarise_cam(
  # Supply your detection data
  detect_data = eh14_detections,
  # Supply your raw image data
  raw_data = eh14_raw,
  # Now specify the time interval you're interested in 
  time_interval = "week",
  # What variable are you interested in?
  variable = "detections",
  # Your desired output format (wide or long) 
  output_format = "wide" 
)

The ultimate pipeline

library(wildRtrax)

Sys.setenv(WT_USERNAME = "*****",
           WT_PASSWORD = "*****")
wt_auth()


projects <- wt_get_download_summary("CAM") %>%
  filter(project == "ABMI Ecosystem Health 2014") %>%
  select(project_id) %>%
  pull()

raw_data <- map_dfr(.x = projects, 
                    .f = ~wt_download_report(.x, "CAM", weather_cols = F, reports = "main")

summarised <- wt_ind_detect(raw_data, 30, "minutes") %>%
  wt_summarise_cam(raw_data, "day", "detections", "long")

And now you can get straight into the science!