Data · Analytics · 2025

Retail Behavior Clustering

An end-to-end pipeline that cleans the Online Retail II dataset, builds RFM features, clusters customers by behavior, and mines basket patterns with association rules.

Customer cluster visualization from the online retail analysis. — Customer clusters from the segmentation workflow.

Problem

Raw transaction logs do not tell you much on their own about who to target or what to stock.

Approach

I turned the transactions into customer-level RFM features, clustered the behavior, and added association-rule mining on top.

Outcome

Clear customer groups and buying patterns that line up with real business decisions.