Curious if to what degree it is possible to cluster epicentres as originating from a single formation/fault? Seems like it is possible that apart of the issue might being able to map the hypocentre of a faults slope, but what is not clear is if that is true, or if faults are really more a collections of formations that vary and at a high-level behave as one, but often independently behave when reviewed on a small scale due to the formations local attributes.
A fault is a crack in the rock that shows movement over time. The movement may have stopped (eg. small faults seen in the field in the UK) or be essentially dead even though the crack still represents a weakness that occasionally shows a small amount of movement (e.g. the Great Glen Fault in Scotland).
Yes it is often possible to map a fault in 3d by plotting the hypocentres in 3d. For example, a good estimate of the fault area that slipped can be found by plotting aftershocks of a large earthquake (although beware of adjoining faults that might see shocks as their stress regime has changed).