Taylor Miller

The examples show too many options. There should be a **run this first.py** that is a simple 10 liner that loads a pre-built dataset, and trains a single model (pick...

Apparently there is an *old* [known issue that RTD doesn't search markdown documents](https://github.com/rtfd/readthedocs.org/issues/1487). A known workaround is to convert .md files to .rst files before they are built by read...

- [ ] mirror the training getting started doc http://healthcareai-py.readthedocs.io/en/master/training/

Documentations could be clearer, for example in deploy.md, it starts off loading a previous model, but it doesn't say where the model should come from or how it's created?

Similar to the intro R videos on healthcare.ai.

This can be general to start and include big outstanding issues and other big picture ideas.

Similar to the R package. The function should find variations across n dimensions for a given outcome and dimensions. This allows the user to surface potential insights in data automatically.