Natural Language Processing of Recovery Group Data

I would like to do an analysis of various timelines of recovery group datasets, it appears there is in many cases a long term evolution of symptoms. Would like to do a word frequency type of analysis over time to understand how it evolves.

Here is the dataset:


Can you elaborate on the consent of the patients?

It is a public group, like many of the public facebook groups, etc so I am not sure if HIPPA etc applies, obviously the names should be anonymized.