FAIR data + privacy metric evaluation algos.
Dataset offers some schema.
Privacy metric requires some set of features.
**Case 1**
dataset schema = required features by privacy metric
**Case 2**
dataset schema $\subset$ required features ->
_2.1:_ feature extraction algo(s) exists, which can computed missing features
_2.2:_ missing features CANNOT be computed form available schema/data
**Case 3**
dataset schema $\supset$ required features -> project schema down into relevant feature space/ leave out uneeded data
**Case 4**
Unknown relationship -> further investigaton needed 
Is this realistic case?