Plot comparisons of smooths

Usage

# S3 method for class 'compare_smooths'
draw(
  object,
  ncol = NULL,
  nrow = NULL,
  guides = "collect",
  select = NULL,
  partial_match = FALSE,
  show = c("models", "difference"),
  pair = NULL,
  ci_level = 0.95,
  surface = c("raster", "tile"),
  contour = TRUE,
  n_contour = NULL,
  continuous_fill = NULL,
  ...
)

Arguments

object: of class "compare_smooths", the result of a call to compare_smooths().
ncol, nrow: numeric; the numbers of rows and columns over which to spread the plots
guides: character; guide handling for the combined plot, passed to patchwork::wrap_plots(). The default is "collect".
select, partial_match: select smooths from the comparison, using the conventions of compare_smooths(). Numeric and logical selections follow the order of unique smooth labels in object.
show: character; "models" (default) overlays one-variable smooths and draws two-variable smooths in side-by-side model facets with fixed axes and a shared fill scale. "difference" draws the estimated difference for selected two-variable smooths, using a fill scale centred on zero.
pair: two model names or integer indices in input order. In difference mode, subtract the second model from the first. With exactly two models, the default uses their input order; with more models, supply pair.
ci_level: numeric; one finite value strictly between 0 and 1, giving the coverage of the pointwise normal intervals for one-variable smooths. The default is 0.95. Use ci_level to change the interval level; coverage and level are not accepted, even when ci_level is also supplied.
surface: character; "raster" or "tile" for surface displays. Irregular numeric grids use rectangles bounded by coordinate midpoints.
contour: logical; add contour lines? These require two numeric axes and at least one grid cell with four finite corner values that are not all equal.
n_contour: numeric; the number of contour bins. Will result in n_contour - 1 contour lines being drawn. See ggplot2::geom_contour().
continuous_fill: a suitable fill scale to be used when plotting continuous variables.
...: arguments passed to patchwork::wrap_plots().

Details

Difference plots match both coordinates and any retained conditioning covariates. Locations missing from either model, and missing estimates, remain missing. Multiple conditioning values are drawn in separate panels. No interval is calculated for the difference: estimates from models fitted to the same observations can be correlated, and their separate covariance matrices do not supply the covariance between models. See compare_smooths() for the scale and centring of the compared smooth contributions.

Label layout

Pass label_control = draw_label_control(...) to draw() to wrap discrete axis labels and legend text, stagger discrete tick labels, or adjust legend and colour-bar layout. Settings apply to each panel, including unwrapped plot lists. See draw_label_control() for the available settings.