Well, I've got years behind me of roadracing motorcycles, although not with any of that fancy datalogging stuff.
If the rider has been talking with the engineers (in my case those people are one and the same) then the engineers would have figured out at what RPM to change to the next gear, and it's basically a matter of shifting slightly above the RPM at which the engine makes peak power such that in the next gear, the engine is making the same power as it did before the shift. This way, on straightaways, the engine is spending its time bracketing the RPM of maximum power. From this point of view the only thing that matters is the maximum power possible. BUT ... it's not that simple in roadracing.
Exiting corners, the thing that matters is driveability, not power. If the bike is underpowered (mine) then it's best to have the engine somewhere near its torque peak mid-corner, and it's best to have *some* room for the rider to play with the range between max torque and max power. (For example, mine makes peak torque around 9500 rpm on the tach and peak power around 13,500 rpm, this works out pretty good on the relatively tight tracks we have here. A peaky engine that makes peak torque very high in the RPM range will make the bike difficult to ride.) The range of RPM and throttle position that the rider uses on track (in my case 8,000 at one tight hairpin corner to 14,000 rpm on straightaways) needs to have no "hiccups", hesitations, or peaks and valleys in the throttle response curves. The way the engine picks up load at slightly-cracked-open throttle (in the useful RPM range) is extremely critical as this condition will occur mid-corner on every corner, and abruptness in load take-up (e.g. caused by stock fuel injection systems having a "deceleration fuel cut" algorithm for emission control) will upset the chassis and can make the rear tire break traction.
I've not driven cars on the track but one would think most of the same principles apply.
