The ... flavour ... I'm getting is that most trips could be considered 'returns' - out and back by public transport.
A couple of exceptions for you:
1. short journeys which are within walking distance but for a serious hill, for example when I was at university, I would catch the bus up the hill but often came back down on foot
2. once one of my brothers had to be in Nottingham, where my father (with a car) already was for some reason. So we put my brother on a train and he came back by car with my father
I'd guess the latter sort of thing is very rare though, your assessment that 'if just one leg (assuming connections are used) cannot be done by public transport, the whole double trip fails.' is one that I agree with, although I would re-word it as 'if the journey can only be done in one direction by public transport, neither direction will be done using public transport'.
Where connections are involved, if one leg fails it might only be that leg that is lost; for example rather than go from A to C, changing at B, pepole might drive to B and use rail from B to C and back again, as long as B to C works in both directions. I think that happens quite a bit with Pembrokeshire services, when part of my family goes to rugby in Cardiff they tend to drive to Port Talbot Parkway and use the train into Cardiff, rather than use a station closer to home. The local
MP▸ (or was it the AM, I forget now) once refered to Port Talbot Parkway as a key station for our area as well.
As for journey metrics, my father recently pointed out that cameras are being installed near junctions on major roads. He thinks they are number-plate recognition systems tracking car movements; that data could be very useful for identifying where there is a big flow of point-to-point journeys which could be captured by a new public transport service.