[Feature addition] Playground Tree

[jdetle]

Issue

The biggest issue facing datreant now is explaining to people working with complicated filesystems what datreant is and why it is useful.

Proposal

Add a Jupyter notebook that takes a large (and fake) sample data set and generates a treant structure from it. This would allow new users to quickly play around with the functionality without looking through the tests.

[dotsdl]

I think this is a good idea. We can make an examples repo that shows how Treants can be used to analyze datasets that are disparate, which is the major thing they do.

[kain88-de]

I think the scipy paper is already a good start for that. If we can replace it with some more real-world data it would be very good.

[orbeckst]

Somewhat related to #88?