Processed over 500,000 transaction records from a UK-based retailer to optimize marketing ROI. Utilized K-Means Clustering to segment customers and performed Cohort Analysis to identify a critical drop-off in user retention during the 3rd month.
- Python: Pandas, NumPy, Scikit-Learn.
- Analysis: RFM (Recency, Frequency, Monetary) Modeling.
- Visualization: Seaborn Heatmaps, Matplotlib.
- Segmentation: Identified 3 distinct customer profiles ("Champions", "Mid-Value", "At-Risk") using unsupervised learning (K-Means).
- Retention Strategy: Cohort analysis revealed a significant retention drop-off (~40% decline) after Month 3, signaling the need for a targeted re-engagement email campaign at the 90-day mark.
- Automation: Built a reusable pipeline to auto-calculate RFM scores for 4,000+ unique customers.
(Retention rates declining over time, highlighting the Month 3 drop-off point.)