Project 03: Mall Customer Segmentation - Discovery Challenge

🎯 Learning Objectives

By completing this project, you will:

• Master K-Means Clustering: Understand and apply the K-Means algorithm
• Apply Data Exploration: Analyze customer data patterns
• Find Optimal Clusters: Use the Elbow Method to determine ideal K value
• Create Visualizations: Plot cluster results with clear labeling
• Business Application: Understand how to segment customers for marketing

Difficulty Level: Intermediate Estimated Time: 2-3 hours Prerequisites: Completed Activity 08 (Clustering)

🛒 Business Context

You own a supermarket mall and have customer data from membership cards:

• Customer ID, Gender, Age
• Annual Income (k$)
• Spending Score (1-100)

Goal: Group customers into segments for targeted marketing strategies!

🚀 Getting Started (See Results in 30 Seconds!)

Option One: Explore the Complete Solution First (Recommended)

1. Open v1-baseline-100percent.ipynb in Google Colab or Jupyter
2. Run all cells (Runtime -> Run all)
3. Observe the complete workflow from data loading to cluster visualization
4. Study the pattern: Explore -> Cluster -> Visualize -> Optimize

Option 2: Jump Straight to the Challenge

1. Open project-03-mall-segmentation.ipynb
2. Run the setup cells to see working examples
3. Complete TODOs in order to build your own clustering model

📋 What's Already Working

✅ Data Download & Loading

• gdown dataset download from Google Drive
• Data loading with pandas
• All necessary library imports

✅ Data Exploration (Phase 2)

• First 5 rows display
• Dataset shape and info
• Missing value detection

✅ K-Means Model Setup (Phase 3)

• Feature extraction code
• KMeans import
• Model creation template

✅ Visualization Template (Phase 4)

• Color scheme for clusters
• Figure size setup
• Label and title formatting

📝 Your Tasks (6 TODOs to Complete)

Phase 3: Build K-Means Model (3 TODOs - Medium)

Expected Inertia: ~23838.25

Phase 4 & 5: Visualization and Optimization (3 TODOs - Medium to Hard)

✅ Success Criteria

You've successfully completed this project when:

• All 6 TODO cells execute without errors
• KMeans model trained with 5 clusters
• Inertia value is approximately 23838.25
• Cluster predictions match expected values
• Scatter plot shows 5 distinct clusters with legend
• Elbow method graph shows decreasing inertia
• Optimal K identified from elbow point

🎓 Key Concepts Applied

One. K-Means Clustering

from sklearn.cluster import KMeans
model = KMeans(n_clusters=5, random_state=42)
model.fit(X)
labels = model.labels_
centers = model.cluster_centers_

2. Elbow Method

inertias = []
for k in range(1, 11):
    model = KMeans(n_clusters=k, random_state=42)
    model.fit(X)
    inertias.append(model.inertia_)
# Plot and find the "elbow" point

3. Cluster Visualization

colors = ['red', 'blue', 'green', 'orange', 'purple']
for i in range(n_clusters):
    cluster_data = X[labels == i]
    plt.scatter(cluster_data[:, 0], cluster_data[:, 1], c=colors[i], label=f'Cluster {i}')

🚀 Extension Challenges

Once you complete all TODOs, try these:

1. Challenge One: Different Features

   • Cluster using Annual Income + Spending Score
   • Compare results with Age + Spending Score

2. Challenge 2: Kneed Library

   • Use kneed.KneeLocator to automatically find optimal K
   • !pip install kneed

3. Challenge 3: 3D Clustering

   • Use Age, Income, and Spending Score together
   • Create 3D scatter plot

4. Challenge 4: Customer Profiling

   • Calculate average values for each cluster
   • Name clusters (e.g., "Young Spenders", "Budget Savers")

💡 Tips for Success

1. Run v1 baseline first: See the complete solution to understand the workflow
2. Use .values: Convert DataFrame columns to numpy arrays for sklearn
3. Set random_state=42: Ensures reproducible results
4. Label your plots: Always add xlabel, ylabel, title, and legend
5. Look for the elbow: The optimal K is where the curve bends most

🐛 Common Issues & Solutions

Issue: "ValueError: could not convert string to float"

Solution: Check that you extracted numerical columns only, not 'Gender' or 'CustomerID'

Issue: "Clusters don't look right"

Solution: Try different random_state values or verify feature extraction

Issue: "Can't find the elbow"

Solution: The elbow is usually around K=3-5 for this dataset

Issue: "Empty cluster warning"

Solution: Reduce n_clusters or try different random_state

📊 Expected Outputs

Phase 3 - Model Results

Inertia of Model with K=5: 23838.24882164186

Cluster Centers:
[[58.44444444, 50.52777778],
 [41.48484848, 37.        ],
 [30.1754386 , 82.35087719],
 [25.4       , 52.68571429],
 [43.28205128, 11.84615385]]

Phase 3 - Customer Predictions

Customer A (Age: 20, Spending: 42) -> Cluster 3
Customer B (Age: 65, Spending: 81) -> Cluster 0
Customer C (Age: 44, Spending: 100) -> Cluster 2
Customer D (Age: 59, Spending: 23) -> Cluster 4

Phase 5 - Optimal K

The elbow point is typically around K=3 or K=4

📚 Additional Resources

• sklearn KMeans Documentation
• Elbow Method Explained
• Customer Segmentation Guide

🎯 Estimated Completion Time

• Fast Track (completed Activity 08): 60-90 minutes
• Standard Pace: 90-120 minutes
• Deep Learning (with challenges): 150-180 minutes

Ready to begin? Open project-03-mall-segmentation.ipynb and start clustering your customers! 🛒