drivendataorg/cookiecutter-data-science ? reverse-engineered prompt

Build me a small Python tool that helps me start a new data science project in a clean, standard way.

I want to run one command and have it ask a few friendly setup questions, then create a ready to use project folder with places for raw data, processed data, notebooks, models, reports, docs, and source code. It should also create a sensible README, a Makefile with a few useful shortcuts, and the basic package files so the project is easy to install and work on later.

Make it feel like a polished starter template for real work, not just a toy example. If it helps, look up current docs online and match the usual best practices for modern Python projects. Also include tests for the template generation so I can trust it works.