Visualizing Clusters in R

Welcome to this lesson on "Visualizing Clusters in R using the Iris Dataset". Previously, we introduced unsupervised learning, focusing on clustering and the K-means clustering algorithm. In this unit, we will visualize the clusters resulting from the K-means algorithm using R’s ggplot2 package. The goal of this lesson is to illustrate the implementation of the K-means clustering algorithm and to demonstrate how to visualize the results using ggplot2, utilizing the classic Iris dataset as an example.

In this lesson, we will use R’s built-in iris dataset. The Iris dataset is a well-known dataset in pattern recognition, consisting of 150 samples from three species of Iris flowers (Iris setosa, Iris virginica, and Iris versicolor), with four features measured for each sample: the length and width of the sepals and petals.

To access the dataset in R, simply use the iris data frame, which is available by default:

# Load the Iris dataset
data(iris)

# View the first few rows of the dataset
head(iris)

Example output of head(iris):

  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
4          4.6         3.1          1.5         0.2  setosa
5          5.0         3.6          1.4         0.2  setosa
6          5.4         3.9          1.7         0.4  setosa

The iris data frame contains five columns: four numeric features (Sepal.Length, Sepal.Width, Petal.Length, Petal.Width) and one categorical column (Species). For clustering, we will use only the numeric features.

Let’s apply the K-means algorithm using R’s built-in kmeans() function and visualize the results using the ggplot2 package. We will focus on plotting the first two features (Sepal.Length and Sepal.Width) for visualization.

Note: Although the Iris dataset contains three species, in this example we use centers = 2 in the K-means algorithm. This is to illustrate how the algorithm groups the data when the number of clusters does not match the number of true species. In practice, you might experiment with different values for centers (such as 3) to see how the clustering changes.

First, we prepare the data and perform K-means clustering:

# Install ggplot2 if not already installed
# install.packages("ggplot2")

library(ggplot2)

# Select only the numeric features for clustering
iris_data <- iris[, 1:4]

# Perform K-means clustering with 2 clusters
set.seed(1)  # For reproducibility
kmeans_result <- kmeans(iris_data, centers = 2, nstart = 10)

# Add cluster assignments to the original data
iris$Cluster <- as.factor(kmeans_result$cluster)

Now, let’s visualize the clusters using ggplot2:

# Create a data frame for cluster centers
centers <- as.data.frame(kmeans_result$centers)
centers$Cluster <- as.factor(1:2)

# Plot using ggplot2
ggplot(iris, aes(x = Sepal.Length, y = Sepal.Width, color = Cluster)) +
  geom_point(size = 2, alpha = 0.7) +
  geom_point(
    data = centers,
    aes(x = Sepal.Length, y = Sepal.Width),
    color = "red",
    shape = 8,
    size = 5,
    stroke = 2
  ) +
  labs(
    title = "Visualizing Clusters in R using Iris Dataset (ggplot2)",
    x = "Sepal Length (cm)",
    y = "Sepal Width (cm)",
    color = "Cluster"
  ) +
  theme_bw() +
  guides(color = guide_legend(override.aes = list(size = 4)))

This code creates a scatter plot of the data points colored by their cluster assignment and marks the cluster centers with large red stars. Check out the plot presented below:

Note: The clustering was performed using all four numeric features, but the plot only shows two dimensions (Sepal.Length vs. Sepal.Width). This means the visualization may not capture the full structure of the clusters as determined by the algorithm in the higher-dimensional space. For a more complete view, you could use dimensionality reduction techniques (like PCA) or create pairwise plots.

Here’s an explanation of the parameters and functions used in the ggplot2 visualization:

• ggplot(iris, aes(x = Sepal.Length, y = Sepal.Width, color = Cluster)) — Initializes the plot, mapping Sepal Length and Sepal Width to the x and y axes, and coloring points by their cluster assignment.

• geom_point(size = 2, alpha = 0.7) — Plots the data points as semi-transparent circles.

• geom_point(data = centers, ...) — Adds the cluster centers to the plot. The shape = 8 parameter uses a star shape, size = 5 increases their size, and color = "red" makes them stand out.

• labs() — Sets the plot title, axis labels, and legend title.

• theme_minimal() — Applies a clean, minimalistic theme to the plot.

• guides(color = guide_legend(override.aes = list(size = 4))) — Makes the legend points larger for better visibility.

Congratulations! In this lesson, you have successfully implemented K-means clustering and visualized clusters using R with the Iris dataset and ggplot2. Practice exercises following this lesson will help you reinforce these skills. Continue practicing and expanding your knowledge of machine learning and data visualization in R. Keep up the great work!