autoplot.ResamplingSpCVKnndm: Visualization Functions for SpCV knndm Method.

Description

Generic S3 plot() and autoplot() (ggplot2) methods to visualize mlr3 spatiotemporal resampling objects.

Usage

# S3 method for ResamplingSpCVKnndm
autoplot(
  object,
  task,
  fold_id = NULL,
  plot_as_grid = TRUE,
  train_color = "#0072B5",
  test_color = "#E18727",
  repeats_id = NULL,
  sample_fold_n = NULL,
  ...
)
# S3 method for ResamplingRepeatedSpCVKnndm
autoplot(
  object,
  task,
  fold_id = NULL,
  repeats_id = 1,
  plot_as_grid = TRUE,
  train_color = "#0072B5",
  test_color = "#E18727",
  sample_fold_n = NULL,
  ...
)
# S3 method for ResamplingSpCVKnndm
plot(x, ...)
# S3 method for ResamplingRepeatedSpCVKnndm
plot(x, ...)

Arguments

object: [Resampling]
mlr3 spatial resampling object of class ResamplingSpCVBlock or ResamplingRepeatedSpCVBlock.
task: [TaskClassifST]/[TaskRegrST]
mlr3 task object.
fold_id: [numeric]
Fold IDs to plot.
plot_as_grid: [logical(1)]
Should a gridded plot using via patchwork be created? If FALSE a list with of ggplot2 objects is returned. Only applies if a numeric vector is passed to argument fold_id.
train_color: [character(1)]
The color to use for the training set observations.
test_color: [character(1)]
The color to use for the test set observations.
repeats_id: [numeric]
Repetition ID to plot.
sample_fold_n: [integer]
Number of points in a random sample stratified over partitions. This argument aims to keep file sizes of resulting plots reasonable and reduce overplotting in dense datasets.
...: Passed to geom_sf(). Helpful for adjusting point sizes and shapes.
x: [Resampling]
mlr3 spatial resampling object. One of class ResamplingSpCVBuffer, ResamplingSpCVBlock, ResamplingSpCVCoords, ResamplingSpCVEnv.

2D vs 3D plotting

This method has both a 2D and a 3D plotting method. The 2D method returns a ggplot with x and y axes representing the spatial coordinates. The 3D method uses plotly to create an interactive 3D plot. Set plot3D = TRUE to use the 3D method.

Note that spatiotemporal datasets usually suffer from overplotting in 2D mode.

Details

This method requires to set argument fold_id and no plot containing all partitions can be created. This is because the method does not make use of all observations but only a subset of them (many observations are left out). Hence, train and test sets of one fold are not re-used in other folds as in other methods and plotting these without a train/test indicator would not make sense.

Examples

Run this code

# \donttest{
if (mlr3misc::require_namespaces(c("CAST", "sf"), quietly = TRUE)) {
  library(mlr3)
  library(mlr3spatiotempcv)
  task = tsk("ecuador")
  points = sf::st_as_sf(task$coordinates(), crs = task$crs, coords = c("x", "y"))
  modeldomain = sf::st_as_sfc(sf::st_bbox(points))

  resampling = rsmp("spcv_knndm",
    folds = 5, modeldomain = modeldomain)
  resampling$instantiate(task)

  autoplot(resampling, task,
    fold_id = 1, size = 0.7) *
    ggplot2::scale_x_continuous(breaks = seq(-79.085, -79.055, 0.01))
}
# }