Plot data in loop • tools4watlas

This article shows how to make and save plots in a loop using a single-core as well as with parallel computing using the package foreach and doFuture. This is convenient for checking the quality of many tags by flicking though them.

Load packages

# packages
library(tools4watlas)
library(ggplot2)
library(foreach)
library(doFuture)

Loop by tag ID

Example of a simple loop by tag ID and using atl_check_tag().

# load example data
data <- data_example

# file path
path <- "./outputs/maps_by_tag/tag_"

# unique ID (here by tag)
id <- unique(data$tag)

# loop to make plots for all
foreach(i = id) %do% {

  # subset data
  data_subset <- data[tag == i]

  # plot and save data
  atl_check_tag(
    data_subset,
    option = "datetime",
    highlight_first = TRUE, highlight_last = TRUE,
    filename = paste0(path, data_subset[1]$species, "_tag_", i)
  )

}

Example to make any type of ggplot and save the data with ragg.

# load example data
data <- data_example

# file path
path <- "./outputs/maps_by_tag/tag_"

# unique ID (here by tag)
id <- unique(data$tag)

# loop to make plots for all
foreach(i = id) %do% {

  # subset data
  data_subset <- data[tag == i]

  # plot data
  p <- ggplot()

  # save
  agg_png(
    filename = paste0(path, i, ".png"),
    width = 3840, height = 2160, units = "px", res = 300
  )
  print(p)
  dev.off()

}

Parallel loop by tag ID

For computing efficieny, we can use parallel computing. We use the same structure, but only replacing %do% with %dofuture%, which is the key advantage of using foreach. The only additional step is setting up parallel processing with registerDoFuture() and plan(multisession).

# load example data
data <- data_example

# file path
path <- "./outputs/maps_by_tag/"

# unique ID (here by tag)
id <- unique(data$tag)

# register cores and backend for parallel processing
registerDoFuture()
plan(multisession)

# loop to make plots for all
foreach(i = id) %dofuture% {

  # subset data
  data_subset <- data[tag == i]

  # plot and save data
  atl_check_tag(
    data_subset,
    option = "datetime",
    highlight_first = TRUE, highlight_last = TRUE,
    filename = paste0(path, data_subset[1]$species, "_tag_", i)
  )

}

# close parallel workers
plan(sequential)

Plot and save data in parallel loop by tag ID and tide ID

# load example data
data <- data_example

# file path
path <- "./outputs/maps_by_tag_and_tide/"

# unique ID combinations
idc <- unique(data[, c("species", "tag", "tideID")])

# register cores and backend for parallel processing
registerDoFuture()
plan(multisession)

# loop to make plots for all
foreach(i = seq_len(nrow(idc))) %dofuture% {

  # subset data
  data_subset <- data[tag == idc$tag[i] & tideID == idc$tideID[i]]

  # plot and save data
  atl_check_tag(
    data_subset,
    option = "datetime",
    highlight_first = TRUE, highlight_last = TRUE,
    filename = paste0(
      path, idc$species[i], "_tag_", idc$tag[i], "_tide_", idc$tideID[i]
    )
  )

}

# close parallel workers
plan(sequential)