Polish Longitudinal Notebook

[eroell]

See comments made by @Zethson here

Notice that comment 4. is addressed with issue #1008

[Zethson]

I killed the old numpy seed usage but please double check that things are still deterministic.
Also made theislab/ehrdata#215 to address another concern.

I will probably still have more comments so please ask for a review when you've addressed all of them.

Discussing an overhaul of the doc structure is out of scope short term so I'm fine with pushing that out a bit more but it's very high up in priorities.