I was not so productive yesterday as I had to run a few errands around the city but I did work more on my knowledge graph spec and how to use it as a recommender system.
I picked the following goals to help me solidify my thought process: 1. Build a Consistent Fitness Routine and 2. Stay Mentally & Socially Engaged.
For each of them, I created flashcards for the first month so I could get an idea of how each action requested of the user could be reflected back to the knowledge graph which in turn reflects back recommendations to the user. Simply thinking about it didn't help, however.
It only made me question if a knowledge graph was the right tool for my recommender system. I have seen how relationships are easily mapped with it so I have to put my doubts away and just push forward.
So I think I will work on actually creating the recommender system with a few goals and I think that will give me a better idea on how to design it.
An experience like this always brings me back to a lesson from my masters program on how strategy consulting is not always a top-down experience. It's a zig-zag between top-down thinking and bottom-up experimentation. Even when the bottom-up experimentation lacks structure, I have found that it helps improve the top-down design.
Anyways, I have to take a test now and gotta run to Shinjuku - I hope I pass.