EDA and RFM Customer Segmentation in E-Commerce (Olist & SQL)
2026-06-19
Comprehensive exploratory data analysis (EDA) and advanced customer segmentation for the Brazilian e-commerce market based on real data from the Olist platform (over 100k orders).
In this project, I combined the performance of a modern in-memory relational engine DuckDB (SQL) with the visualization capabilities of Python (Seaborn & Matplotlib) libraries. The main goal was to examine data referential integrity, identify bottlenecks in logistics operations, and analyze their direct impact on consumer satisfaction and ratings.
The culmination of the analysis is a custom RFM (Recency, Frequency, Monetary) segmentation model implemented in pure SQL, which divides the customer base into strategic business groups (e.g., Champions, Loyal Customers, At Risk), enabling tailored marketing actions. The project perfectly demonstrates a business-oriented approach and advanced data modeling skills using SQL and Python.
The project was written in Polish.