Re: Why do Mazda fake the data?
I've worked with plenty of automotive data. Fake data such as engine "actual" torque output is a filtered calculation based on fuel rate (measured, maybe; likely estimated from injector pattern), boost pressure (actual*), engine speed, etc. Engine supplier echoed John Arthur: users are idiots and don't need actual data, as if engines had actual torque meters on the crankshaft! (*Boost pressure at intake manifold really was actual and you could see the surging every upshift.)
I made up nice graphical dashboards for a chassis dynamometer test cell running off CAN data -- great to see a general picture of what's going on in the vehicle (allowing driver-less operation), but worthless to diagnose an actual problem. It is the CAN-based "trouble codes" (visible to user as "idiot lamps" or "flash codes", or via OBDC readers) that one needs to look for.
I've made plenty of other CAN dashboards for various projects. The usefulness is in changing each one to fit the needs of the test of the day: What data will (quickly and reliably) tell me if item X is working properly or not? It helps when the data you want is actually available on the CAN bus.