To be honest, it’s very hard to tell! It looks like a co-occurrence matrix of some kind?
I understand it would probably be a large undertaking (involving talking to all the people with a bit of knowlege and reading the existing endpoints that query the database), but I think you’ll struggle the least if you took the time to try to map out the entire model and domain with new, meaningful names and the inter-relationships between them. I imagine it might take a few weeks to map out and organise a good schema but I think the payoff would be worthwhile!
That being said, if you just want to proceed naiively you can try to just insert an entity type
smem_uncommitted_spread with the attributes
However, let’s say
num_appearances_i_j is encoding a matrix where the value at index
[i,j] represents a number attached to an edge/the number of edge between some data point
j, then you would actually want to move to a real graph structure instead of a matrix, creating a relation between concept
j, with an attribute representing the number of occurrences. Which takes me back to actually trying to understand what the data represents and re-modeling it properly!
This might be a lot to digest, and I probably got a lot of guesses wrong, but I hope it helps a bit!