tidiers {ordr}

Tidiers for 'tbl_ord' objects

Description

These functions return tibbles that summarize an object of class 'tbl_ord'. tidy() output contains one row per artificial coordinate and glance() output contains one row for the whole ordination.

Usage

## S3 method for class 'tbl_ord'
tidy(x, ...)

## S3 method for class 'tbl_ord'
glance(x, ...)

## S3 method for class 'tbl_ord'
fortify(model, data, ..., .matrix = "dims", elements = "all")

Arguments

x, model

An object of class 'tbl_ord'.

...

Additional arguments allowed by generics; currently ignored.

data

Passed to generic methods; currently ignored.

.matrix

A character string partially matched (lowercase) to several indicators for one or both matrices in a matrix decomposition used for ordination. The standard values are "rows", "cols", and "dims" (for both).

elements

Character vector; which elements of each factor for which to render graphical elements. One of "all" (the default), "active", or any supplementary element type defined by the specific class methods (e.g. "score" for 'factanal', 'lda_ord', and 'cancord_ord' and "intraset" and "interset" for 'cancor_ord').

Details

Three generics popularized by the ggplot2 and broom packages make use of the augmentation methods:

• The generics::tidy() method summarizes information about model components, which here are the artificial coordinates created by ordinations. The output can be passed to ggplot2::ggplot() to generate scree plots. The returned columns are

  • name: (the name of) the coordinate

  • other columns extracted from the model, usually a single additional column of the singular or eigen values

  • inertia: the multidimensional variance

  • prop_var: the proportion of inertia

  • quality: the cumulative proportion of variance

• The generics::glance() method reports information about the entire model, here always treated as one of a broader class of ordination models. The returned columns are

  • rank: the rank of the ordination model, i.e. the number of ordinates

  • n.row,n.col: the dimensions of the decomposed matrix

  • inertia: the total inertia in the ordination

  • ⁠prop.var.*⁠: the proportion of variance in the first 2 ordinates

  • class: the class of the wrapped model object

• The ggplot2::fortify() method augments and collapses row and/or column data, depending on .matrix and .element, into a single tibble, in preparation for ggplot2::ggplot(). Its output resembles that of generics::augment(), though rows in the output may correspond to rows, columns, or both of the original data. If .matrix is passed "rows", "cols", or "dims" (for both), then fortify() returns a tibble whose fields are obtained, in order, via ⁠get_*()⁠, ⁠recover_aug_*()⁠, and ⁠annotation_*()⁠.

The tibble is assigned a "coordinates" attribute whose value is obtained via get_coord(). This facilitates some downstream functionality that relies on more than those coordinates used as position aesthetics in a biplot, in particular stat_spantree().

Value

A tibble.

See Also

augmentation methods that must interface with tidiers.

Examples

# illustrative ordination: PCA of iris data
iris_pca <- ordinate(iris, ~ prcomp(., center = TRUE, scale. = TRUE), seq(4L))

# use `tidy()` to summarize distribution of inertia
tidy(iris_pca)
# this facilitates scree plots
tidy(iris_pca) %>%
  ggplot(aes(x = name, y = prop_var)) +
  geom_col() +
  scale_y_continuous(labels = scales::percent) +
  labs(x = NULL, y = "Proportion of variance")

# use `fortify()` to prepare either matrix factor for `ggplot()`
fortify(iris_pca, .matrix = "V") %>%
  ggplot(aes(x = name, y = PC1)) +
  geom_col() +
  coord_flip() +
  labs(x = "Measurement")
iris_pca %>%
  fortify(.matrix = "U") %>%
  ggplot(aes(x = PC1, fill = Species)) +
  geom_histogram() +
  labs(y = NULL)
# ... or to prepare both for `ggbiplot()`
fortify(iris_pca)

# use `glance()` to summarize the model as an ordination
glance(iris_pca)
# this enables comparisons to other models
rbind(
  glance(ordinate(subset(iris, Species == "setosa"), prcomp, seq(4L))),
  glance(ordinate(subset(iris, Species == "versicolor"), prcomp, seq(4L))),
  glance(ordinate(subset(iris, Species == "virginica"), prcomp, seq(4L)))
)