Clustering Antarctic Penguin Species

# Import Required Packages
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.cluster import KMeans
from sklearn.metrics import silhouette_score
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.manifold import TSNE

# Loading and examining the dataset
penguins_df = pd.read_csv("penguins.csv")
df = penguins_df.copy()
penguins_df.head()

# Checking for missing data
df.isna().sum()

# The sex column needs to be encoded
encoder = LabelEncoder()
df['sex'] = encoder.fit_transform(df['sex'])
df.head()

# Checking for outliers using box plots and saving the chart
plt.figure(figsize=(15, 10))

# List of columns to check for outliers
columns = ['culmen_length_mm', 'culmen_depth_mm', 'flipper_length_mm', 'body_mass_g']

for i, column in enumerate(columns, 1):
    plt.subplot(2, 2, i)
    sns.boxplot(x=df[column])
    plt.title(f'Box plot of {column}')

plt.tight_layout()
plt.savefig('outliers_boxplot.png')
plt.show()

# Scaling the data
scaler = StandardScaler()
df_scaled = scaler.fit_transform(df)

# Determining the best k clusters
scores = []
for k in np.arange(1, 11):
    model = KMeans(n_clusters= k, random_state=42)
    model.fit(df_scaled)
    scores.append(model.inertia_)
    
# Plotting the elbow plot
plt.figure(figsize=(10, 6))
plt.plot(np.arange(1, 11), scores, marker='o')
plt.xlabel('Number of clusters (k)')
plt.ylabel('Inertia')
plt.title('Elbow Method For Optimal k')
plt.savefig('elbow_plot.png')
plt.show()

# Determine the best k using the silhouette score
silhouette_scores = []

for k in np.arange(2, 11):  # Start from 2 instead of 1
    model = KMeans(n_clusters=k, random_state=42)
    labels = model.fit_predict(df_scaled)
    silhouette_avg = silhouette_score(df_scaled, labels)
    silhouette_scores.append(silhouette_avg)

# Plotting the silhouette scores
plt.figure(figsize=(10, 6))
plt.plot(np.arange(2, 11), silhouette_scores, marker='o')  # Adjust x-axis range
plt.title('Silhouette Score for Different k Values')
plt.xlabel('Number of clusters (k)')
plt.ylabel('Silhouette Score')
plt.xticks(np.arange(2, 11))
plt.grid(True)
plt.savefig('silhouette_scores_plot.png')
plt.show()

# Final model with 5 clusters
model = KMeans(n_clusters=5, random_state=42)
labels = model.fit_predict(df_scaled)

# Visualizing the final clusters
plt.figure(figsize=(10, 6))
sns.scatterplot(x=df_scaled[:, 0], y=df_scaled[:, 1], hue=labels, palette='viridis', legend='full')
plt.scatter(model.cluster_centers_[:, 0], model.cluster_centers_[:, 1], s=300, c='red', label='Centroids')
plt.title('Final KMeans Clustering with k=5')
plt.xlabel('Feature 1')
plt.ylabel('Feature 2')
plt.legend()
plt.savefig('final_kmeans_clustering.png')
plt.show()

# Calculating Silhouette Score
silhouette_avg = silhouette_score(df_scaled, labels)
print(f"Silhouette Score: {silhouette_avg:.3f}")

stat_penguins = df.drop('sex', axis=1)
stat_penguins["Cluster"] = labels
stat_penguins = round(stat_penguins.groupby("Cluster").mean(),2)

# Display the final DataFrame
stat_penguins.head()

# Applying t-SNE for dimensionality reduction
tsne = TSNE(n_components=3, random_state=42)
df_tsne = tsne.fit_transform(df_scaled)

# Visualizing the clusters with t-SNE
plt.figure(figsize=(10, 6))
sns.scatterplot(x=df_tsne[:, 0], y=df_tsne[:, 1], hue=labels, palette='viridis', legend='full')
plt.title('t-SNE enhanced Clusters Visualization')
plt.xlabel('t-SNE Feature 1')
plt.ylabel('t-SNE Feature 2')
plt.legend()
plt.savefig('t-SNE_clusters_visualization.png')
plt.show()

# Get the centroid coordinates
centroids = model.cluster_centers_

abb= scaler.inverse_transform(model.cluster_centers_)

# Convert to DataFrame
centroid_df = pd.DataFrame(centroids[:,:4], columns= penguins_df.drop('sex', axis=1).columns) 
centroid1_df = pd.DataFrame(abb[:,:4], columns= penguins_df.drop('sex', axis=1).columns)  

print(centroid_df)
print(centroid1_df)