[Commerce] Build Etsy order-to-item analytics model

[Nickel85]

Goal

Create a joined analytics model that combines Etsy sold orders and sold order items so recommendations can use order economics and listing-level demand together.

Inputs

• Imported sold orders CSV records.
• Imported sold order item CSV records.
• Order ID, Listing ID, Transaction ID, Item Name, Quantity, Price, Shipping, Fees, Discounts, Net Order Amount, Variations.

Acceptance Criteria

• Joins order-level and item-level records by Order ID.
• Computes listing-level metrics: units sold, revenue, discount usage, average item price, average order value contribution, and repeat order frequency where data supports it.
• Computes order-level metrics: item count, gross total, discount total, shipping paid, processing fees, net amount, and effective margin proxy.
• Identifies digital vs physical listing signals from listing type, item name, shipping, and variations.
• Emits structured analytics facts that can feed LLM recommendations without exposing unnecessary buyer PII.
• Includes tests for single-item orders, multi-item orders, coupons, shipping, missing shipped date, and duplicate imports.