Before I start reworking the structure into a revised data model, it would be very helpful to see a mockup of the tables/analyses/maps you want to create from it. Per my post below, I typically do this type of storyboarding before structuring my data models so that I can be sure the structure I come up with will support the analyses and visuals I ultimately want. While there are some very clearly wrong ways to model your data, there is not always one right or best way to do it - “right or best” are ultimately relative to what you want to do with the data - particularly in this case where there is not a typical fact table and some of the dimensions optimally could be combined or split out separately, depending on the end goals.
Once those objectives specified more explicitly, the “correct” data model usually becomes clear pretty quickly. I suspect you will see this dynamic play out in Data Challenge #3, where some participants will unpivot the interview data while others will maintain it in pivoted form, depending on how they want to analyze the data.
I hope this is helpful to you, even if not the clear-cut answer you were probably looking for.
- Brian