This may come down to personal preference, but I think it is best to have a folder containing all .py files and separate folders for images and data.

Moving forward, this would allow instructors to show students how to organize a module of .py files into submodules and how to import from those submodules.

It might make sense to keep to the current structure for mod 1, create a folder for all .py files in mod 2, create submodules in mod3, etc.

Once students have been introduced to this repo organization, I think either in mod 2 or 3 students should be walked through the process of creating a project repo from scratch.

Because the majority of their work, including their end of mod projects, involve forking a learn repo, I've found that students often do not know how to create a project from scratch. (ie create an empty github repo and push to it)

Personal preference is to work hub into the workflow to encourage students to use the command line. Also aliasing, but that isn't as important

Thinking about this, it might make sense to create a PyPI package that would initiate a vanilla data science repository structure. I would love that for my own life, would probably lead to more organized student repos, and would be a nice tool for students after graduating.

I like the inclusion of the directory tree in the README template. Linking to a simple tutorial on how to print this out would save students some time. This short tutorial covers installation and basic use for macs and this tutorial seems decent for windows, though I have no experience using windows.