Skip to contents

Cluster scatter plot

Source: R/clustering.R
cluster_plot.Rd

Runs k-means clustering on the numeric columns of a data frame (or uses a supplied cluster assignment) and plots the observations on the first two principal components, coloured by cluster, with optional convex hulls and labelled centroids. Projecting onto principal components means the plot works for any number of variables.

Usage

cluster_plot(
  data,
  cols = NULL,
  k = 3,
  clusters = NULL,
  scale = TRUE,
  suggest_k = NULL,
  k_range = 2:8,
  hulls = TRUE,
  label_centers = TRUE,
  point_alpha = 0.8,
  palette = NULL,
  title = NULL,
  seed = NULL,
  nstart = 10,
  iter.max = 10
)

Arguments

data

A data frame.

cols

Numeric columns to cluster on (default: all numeric).

k

Number of clusters for k-means (ignored if clusters is supplied, or overridden when suggest_k chooses a value).

clusters

Optional vector of cluster assignments (e.g. from stats::kmeans() or stats::cutree()); use this to plot a clustering you computed yourself. Hierarchical or PAM assignments (such as stats::cutree() output matching dendrogram_plot()) are accepted. Must have exactly one entry per row of data (rows with missing values in cols are dropped from both before plotting).

scale

Whether to scale variables to unit variance before clustering and the PCA.

suggest_k

Optionally choose k automatically with a cluster-quality diagnostic instead of using the supplied k. Either TRUE (uses the average-silhouette criterion), a string naming the criterion ("silhouette", "wss" or "gap"; see k_diagnostic()), or NULL (the default) to leave k unchanged. Ignored when clusters is supplied.

k_range

Candidate values of k searched when suggest_k is set.

hulls

Whether to draw a shaded convex hull around each cluster.

label_centers

Whether to label the cluster centroids.

point_alpha

Point transparency.

palette

Colours for the clusters; defaults to depictr_palette().

title

Plot title.

seed

Optional integer seed for reproducible k-means (ignored when clusters is supplied). Default NULL leaves the RNG state untouched.

nstart

Number of random starts for stats::kmeans().

iter.max

Maximum iterations for stats::kmeans().

Value

A ggplot2::ggplot object. When suggest_k is used, the chosen k and the full diagnostic are attached as the attribute "k_diagnostic".

See also

k_diagnostic() and silhouette_plot() for cluster-quality diagnostics.

Examples

cluster_plot(crop_yield, cols = c("rainfall", "fertiliser", "soil_ph",
                                  "yield"), k = 3, seed = 1)

# Let a silhouette diagnostic choose k:
cluster_plot(crop_yield, cols = c("rainfall", "fertiliser", "soil_ph",
                                  "yield"), suggest_k = TRUE, k_range = 2:6,
             seed = 1)