plot_clusters: Plot clusters

In pbs-assess/gfranges: Groundfish Ranges

Description

Plot clusters

Usage

plot_clusters(
  model,
  data = NULL,
  colour_vector = NULL,
  text_label = NULL,
  shape_by_group = FALSE,
  colour_label = "model",
  ...
)

Arguments

model	An object of class "partition" created by the functions [cluster::pam()] or [stats::kmeans()].
data	The original data frame used in the clustering.
colour_vector	A vector of character or factors to colour the points by.
colour_label	A label for the colour legend.
...	Other arguments to pass2 [factoextra::fviz_cluster()].

Value

A ggplot2 object.

Examples

df <- haddock_mod %>%
  dplyr::select(fo_median, ffmsy_median, bbmsy_median) %>%
  scale()

# Evaluate the number of clusters
factoextra::fviz_nbclust(df, kmeans, method = "wss")
factoextra::fviz_nbclust(df, kmeans, method = "silhouette")
factoextra::fviz_nbclust(df, kmeans, method = "gap_stat")

# Illustrate an example with 2 clusters:
m <- kmeans(df, centers = 2L)

plot_clusters(m,
  data = df, colour_vector = haddock_mod$fmodel,
  colour_label = "F model"
)

# The default is to plot the two first principal components as x and y.
# Instead we can pick specific axes to show:
plot_clusters(m,
  data = df, colour_vector = haddock_mod$fmodel,
  colour_label = "F model",
  choose.vars = c("bbmsy_median", "ffmsy_median")
)

# Or with the more robust cluster::pam() algorithm:
m <- cluster::pam(df, k = 4L)
plot_clusters(m,
  data = df, colour_vector = haddock_mod$fmodel,
  colour_label = "F model"
)

pbs-assess/gfranges documentation built on Dec. 13, 2021, 4:50 p.m.