facebookexperimental/Robyn ? reverse-engineered prompt

Build me an open source marketing mix modeling package that helps me understand which channels are actually driving sales and how to split budget better.

I want it to take a dataset with media spend, sales, and other factors, then fit a model that can handle adstock, saturation, trend, and seasonality, and also suggest the best budget allocation. Please include a simple end to end demo with sample data so I can run it right away, plus clear notebook style tutorials that show the full flow from loading data to training, checking results, and generating easy to read output.

It would be great if it works in both R and Python, with the Python version set up as a usable beta, and if there is a small web site or app for documentation and examples. Keep it practical, reproducible, and easy for someone with limited technical background to follow. If you need to look up current docs online, do that.