K-Means Clustering Analysis

gistfile1.txt

import plotly.graph_objects as go
import plotly.subplots as sp

# Example data
k_values = list(range(1, 11))
wss_values = [1000, 800, 650, 500, 450, 400, 350, 330, 320, 310]  # Elbow plot data
silhouette_scores = [
    0.1,
    0.3,
    0.45,
    0.52,
    0.51,
    0.50,
    0.48,
    0.47,
    0.46,
    0.45,
]  # Silhouette scores data
dimensions = list(range(1, 11))
explained_variance_ratio = [
    0.1,
    0.18,
    0.25,
    0.30,
    0.35,
    0.38,
    0.40,
    0.42,
    0.43,
    0.44,
]  # Explained variance data

# Create subplots
fig = sp.make_subplots(
    rows=1,
    cols=3,
    subplot_titles=(
        "Elbow Method for Optimal k",
        "Silhouette Score vs. Number of Dimensions",
        "Variance Explained vs. Dimensions",
    ),
)

# Elbow Method Plot
fig.add_trace(
    go.Scatter(x=k_values, y=wss_values, mode="lines+markers", name="WSS"), row=1, col=1
)
fig.add_vline(x=4, line_dash="dash", line_color="red", row=1, col=1)
fig.add_annotation(
    x=4,
    y=500,
    text="Elbow Point (k=3)<br>WSS not reduced<br>significantly after this point",
    showarrow=True,
    arrowhead=2,
    ax=100,
    ay=-100,
    row=1,
    col=1,
)

# Silhouette Score vs. Number of Dimensions
fig.add_trace(
    go.Scatter(
        x=dimensions, y=silhouette_scores, mode="lines+markers", name="Silhouette Score"
    ),
    row=1,
    col=2,
)
fig.add_vline(x=4, line_dash="dash", line_color="green", row=1, col=2)
fig.add_annotation(
    x=4,
    y=0.52,
    text="Highest Silhouette Score<br>Indicates best clustering",
    showarrow=True,
    arrowhead=2,
    ax=-75,
    ay=-75,
    row=1,
    col=2,
)

# Variance vs. Dimensions
fig.add_trace(
    go.Scatter(
        x=dimensions,
        y=explained_variance_ratio,
        mode="lines+markers",
        name="Explained Variance Ratio",
    ),
    row=1,
    col=3,
)
fig.add_vline(x=4, line_dash="dash", line_color="blue", row=1, col=3)
fig.add_annotation(
    x=4,
    y=0.30,
    text="Variance Levels Off<br>Explained variance<br>increase is marginal",
    showarrow=True,
    arrowhead=2,
    ax=-75,
    ay=-75,
    row=1,
    col=3,
)

# Update layout
fig.update_layout(
    height=600,
    width=1800,
    title_text="Analysis Plots for K-Means Clustering",
    showlegend=False,
)
fig.update_xaxes(title_text="Number of Clusters (k)", row=1, col=1)
fig.update_yaxes(title_text="Within-Cluster Sum of Squares (WSS)", row=1, col=1)
fig.update_xaxes(title_text="Number of Dimensions", row=1, col=2)
fig.update_yaxes(title_text="Silhouette Score", row=1, col=2)
fig.update_xaxes(title_text="Number of Dimensions", row=1, col=3)
fig.update_yaxes(title_text="Explained Variance Ratio", row=1, col=3)

# Show the plot
fig.show()