🏋️‍♀️ This is the final challenge of the week (you have until the end of the Communicate topic to complete it).

Before diving into it, take time to finish the challenge 02 on sellers analysis (Liner Regression topic).

Problem

❓ How many underperforming sellers should Olist remove to improve its profit, given that it has:

• some revenues per sellers per months
• some revenues per orders
• some reputation costs (estimated) per bad reviews
• some operational costs of IT system that grows with number of order items, but not linearly (scale effects)

💡 Hints

- You may want to compute the review cost of each order first, then sum these costs by seller.

✏️ Your turn!

👉 Open the ceo_request.ipynb notebook and start from there.

• We'll start from a new Notebook
• Refrain from re-using from previous notebooks - they were made for investigation only
• All your re-usable logic is coded in olist/*.py scripts