K-Means Clustering Visualization in R: Step By Step Guide

k-means-clustering-visualization-in-r-plot-k-means-in-r-1.png

02 Jun

K-Means Clustering Visualization in R: Step By Step Guide

This article provides examples of codes for K-means clustering visualization in R using the factoextra and the ggpubr R packages. You can learn more about the k-means algorithm by reading the following blog post: K-means clustering in R: Step by Step Practical Guide.

Contents:

Required R packages
Data preparation
K-means clustering calculation example
Plot k-means
- Using the factoextra R package
- Using the ggpubr R package
Conclusion

Required R packages

ggpubr: creates publication ready plots.
factoextra: Extract and Visualize the Results of Multivariate Data Analyses.

library(ggpubr)
library(factoextra)

Data preparation

data("iris")
df <- iris
head(df, 3)

##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1          5.1         3.5          1.4         0.2  setosa
## 2          4.9         3.0          1.4         0.2  setosa
## 3          4.7         3.2          1.3         0.2  setosa

K-means clustering calculation example

Removing the 5th column (Species) and scale the data to make variables comparable
Calculate k-means clustering using k = 3. As the final result of k-means clustering result is sensitive to the random starting assignments, we specify nstart = 25. This means that R will try 25 different random starting assignments and then select the best results.

# Compute k-means with k = 3
set.seed(123)
res.km <- kmeans(scale(df[, -5]), 3, nstart = 25)
# K-means clusters showing the group of each individuals
res.km$cluster

##   [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 3 2 2 2 3 2 2 2
##  [61] 2 2 2 2 2 3 2 2 2 2 3 2 2 2 2 3 3 3 2 2 2 2 2 2 2 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 3 2 3 3 3 3 2 3 3 3 3 3 3 2 2 3 3 3 3 2
## [121] 3 2 3 2 3 3 2 3 3 3 3 3 3 2 2 3 3 3 2 3 3 3 2 3 3 3 2 3 3 2

Plot k-means

Using the factoextra R package

The function fviz_cluster() [factoextra package] can be used to easily visualize k-means clusters. It takes k-means results and the original data as arguments. In the resulting plot, observations are represented by points, using principal components if the number of variables is greater than 2. It’s also possible to draw concentration ellipse around each cluster.

fviz_cluster(res.km, data = df[, -5],
             palette = c("#2E9FDF", "#00AFBB", "#E7B800"), 
             geom = "point",
             ellipse.type = "convex", 
             ggtheme = theme_bw()
             )

Using the ggpubr R package

If you want to adapt the k-means clustering plot, you can follow the steps below:

Compute principal component analysis (PCA) to reduce the data into small dimensions for visualization
Use the ggscatter() R function [in ggpubr] or ggplot2 function to visualize the clusters

Compute PCA and extract individual coordinates

# Dimension reduction using PCA
res.pca <- prcomp(df[, -5],  scale = TRUE)
# Coordinates of individuals
ind.coord <- as.data.frame(get_pca_ind(res.pca)$coord)
# Add clusters obtained using the K-means algorithm
ind.coord$cluster <- factor(res.km$cluster)
# Add Species groups from the original data sett
ind.coord$Species <- df$Species
# Data inspection
head(ind.coord)

##   Dim.1  Dim.2   Dim.3    Dim.4 cluster Species
## 1 -2.26 -0.478  0.1273  0.02409       1  setosa
## 2 -2.07  0.672  0.2338  0.10266       1  setosa
## 3 -2.36  0.341 -0.0441  0.02828       1  setosa
## 4 -2.29  0.595 -0.0910 -0.06574       1  setosa
## 5 -2.38 -0.645 -0.0157 -0.03580       1  setosa
## 6 -2.07 -1.484 -0.0269  0.00659       1  setosa

# Percentage of variance explained by dimensions
eigenvalue <- round(get_eigenvalue(res.pca), 1)
variance.percent <- eigenvalue$variance.percent
head(eigenvalue)

##       eigenvalue variance.percent cumulative.variance.percent
## Dim.1        2.9             73.0                        73.0
## Dim.2        0.9             22.9                        95.8
## Dim.3        0.1              3.7                        99.5
## Dim.4        0.0              0.5                       100.0

Visualize k-means clusters

Color individuals according to the cluster groups
Change point shapes according to the Species groups (ground truth of grouping)
Add concentration ellipses
Add cluster centroid using the stat_mean() [ggpubr] R function

ggscatter(
  ind.coord, x = "Dim.1", y = "Dim.2", 
  color = "cluster", palette = "npg", ellipse = TRUE, ellipse.type = "convex",
  shape = "Species", size = 1.5,  legend = "right", ggtheme = theme_bw(),
  xlab = paste0("Dim 1 (", variance.percent[1], "% )" ),
  ylab = paste0("Dim 2 (", variance.percent[2], "% )" )
) +
  stat_mean(aes(color = cluster), size = 4)

Conclusion

This article provides examples of codes for K-means clustering visualization in R using the factoextra and the ggpubr R packages. Read more about the k-means clustering algorithm at: K-means clustering in R: Step by Step Practical Guide.