Olist E-Commerce Analysis

Context

Olist’s Kaggle dataset is one of the most complete public e-commerce datasets available. But most analyses out there stop at basic EDA. I wanted to build an analysis that could actually drive business decisions.

Approach

• Built reproducible data marts from raw data
• Deep EDA: purchase patterns, customer segmentation, review analysis
• Delivery performance analysis and its impact on satisfaction
• Identified product categories with best margin and volume
• Summarized findings in executive insight format

Key Findings

• Delivery time is the number one factor affecting review scores, not price
• There’s a sweet spot in delivery time where review scores drop drastically
• Sellers with fast response times have significantly higher retention rates
• Some product categories have high margins but low volume, perfect for niche strategy

What’s Different

I intentionally built this with the mindset of “if I present this to a CEO, they should be able to take action from this slide.” Not just pretty charts, but every visualization comes with an actionable recommendation.