Project Showcase
Highlight of my featured works
Classifying Purchase Frequency
An unsupervised learning project applying clustering to classify customers by purchase frequency. Includes scaling, encoding, PCA, and evaluation of clustering quality.
🎯 What I Learned
- Applied K-Means clustering with scaling and encoding
- Evaluated clusters with Silhouette and Davies–Bouldin scores
- Visualized clusters using PCA for dimensionality reduction
🛠 Tech Stack
- Python, Pandas, NumPy
- Matplotlib, Seaborn, Plotly
- Statsmodels
- Scikit-learn
📌 Evaluation / Next Step
K-Means produced well-separated clusters with good Silhouette scores after preprocessing. However, K-Means assumes spherical clusters, which may limit its performance on complex data distributions. Next steps include exploring DBSCAN, hierarchical clustering, and ensemble clustering methods.