Anaghatotade/ChainSight ? reverse-engineered prompt

Build me a retail demand and inventory analytics project using the Kaggle M5 Walmart sales data.

I want a notebook driven workflow that takes the raw sales, calendar, and price files, cleans and validates them, then explores the data, calculates revenue and demand KPIs, and shows clear charts and tables that explain what matters for the business. Please include ABC style item analysis, category and store level insights, seasonality, weekday and event effects, SNAP related demand patterns, price and demand relationships, and a section for intermittency and variability.

Then add weekly demand forecasting with simple baselines first, and compare them to a small neural network so I can see whether the extra complexity is worth it. Use a time based split, report forecast error clearly, and make the final recommendation easy to understand.

If anything needs current docs or dataset details, look them up online. Keep the results reproducible, save outputs, and make the final writeup feel like a polished business report.