Introduction

This is a simplified version of a problem that we have encountered and solved at float.

Situation: A customer's 3rd party integration has broken. It is no longer possible to see through Shopify what the end-customer is returning; only the amount they're returning for.

Problem:
To be able to deliver our analytical insights we need to know what items has been returned.

Data
sales.csv is a sample dataset illustrating this broken integration. It contains the following columns:

• order_id (int): The id of the order
• item_id (int): The line item id of the order
• product_id (int): The id of the product that was sold
• product_title (str): The title of the product
• sales (int): The sales value of the item

returns.json is a json file where each key and corresponding value is the order_id and the amount the end-customer has returned for The returned amount corresponds to some combination of the items that was sold on the order. The problem is just: We don't know which.

What you should do
Write up a solution that figures out what products has been returned for each order. Test and document your solution based on the sample dataset: sales.csv and returns.json

Solution rules:

• The solution must be written in Python
• The ideal solution is well-structured and well-documented
• You can use any external libraries of your choice
• Bonus points: Think about the limitations of your solution and how well it scales.