torokmg/m5-time-series ? reverse-engineered prompt

Build me a Jupyter notebook project for the M5 Walmart sales forecasting competition.

I want it to walk through the whole process from getting the data, cleaning it, exploring it, and then building a few forecasting models so I can compare how they do. Please include a simple analysis of the business questions too, like which state and store bring in the most revenue, which product category sells best, and which items are strongest in each store. Then make a 28 day sales forecast for the items and save the predictions in a format that could be used for submission.

Use clear charts and tables so the results are easy to understand, and keep the notebooks organized so someone can follow the story from raw data to final recommendations. If you need to look up current docs or best practices for the libraries, feel free to do that.