Originally Posted by Gabe15

This is an interesting idea, but it's actually quite difficult to sort all of the relevant data, or even to know which data points are relevant or not, and which change by environment and age of the plant, and which do not. There is also the very real situation that routinely arises where two different people are observing the same plant and using the same set of identification characters, but arrive at different results.

On a larger scale, this is an on going effort involving many different researchers using a variety of techniques to answer the simple, but sometimes very difficult question of: how can a given variety be reliably identified? Some of it would transfer well to the hobby scene for dedicated enthusiasts willing to put some time to study it all.

Here are some examples and starting points.

This is the standard set of descriptors used for Musa characterization (there are modified versions for specific subgroups, notably Plantains and Mutika/Lujugira).
https://www.bioversityinternational....s/pdfs/326.pdf

Here is an example of an accession (in this case, a Silk cultivar) in the MGIS which has photos and ratings for the minimum set of descriptors which is a subset of the whole set linked above.
https://www.crop-diversity.org/mgis/...on/01BEL084769