Currently the dishes for the initial quiz are picked random w.r.t to the users hard filter (categories, allergies, additives). But it could be the case that highly similar dishes are presented to the user which can lead to a bad initial rating distribution. It would be nice to consider three dishes with the highest distance between each other (based on the features categories, allergies and additives).
Currently the dishes for the initial quiz are picked random w.r.t to the users hard filter (categories, allergies, additives). But it could be the case that highly similar dishes are presented to the user which can lead to a bad initial rating distribution. It would be nice to consider three dishes with the highest distance between each other (based on the features categories, allergies and additives).