I would do "process control" analysis of the bolt circle on the X and Y coordinates from one common "processing orientation" for the sample.
It is very important that the control is analyzed on pattern in a common process orientation if there are multiple machines segregate them as well.
You would have to convert your basic (if they are detailed like you say).
(R)_ (A)_ (X)_ (Y)
(57.15)_ (0)_ (0)_ (57.15)
(57.15)_ (72)_ (17.66032)_ (54.35288)
(57.15)_ (144)_ (-46.2353)_ (33.59193)
(57.15)_ (216)_ (-46.2353)_ (-33.59193)
(57.15)_ (288)_ (17.66032)_ (-54.32588)
Then if the data appears to be "in control"... you should balance the rotational devation among the five bolt holes, re-compute your rotated X and y coordinates, then go ahead and figure your diameter of position deviation for each, and finally estimate the conformance capability.
Your position deviation distribution is likely going to be skewed toward zero position deviation so you will have to use a skewed distribution model to better predict the probability of a defect.
BTW... It may be more appropriate to use the BCD acronym "bolt circle diameter" instead of P.C.D.
I doubt that there are variable tolerance modifiers on the wheel studs since the "center hub pilot diameter" and the beveled lug nuts both contend to constrain the wheel location regardless of hole clearances... but if there are you will either have to ignore the MMC tolerance or do considerably more reading on this forum to figure out how to accomplish the analysis.
Either way the attached presentation should help you to understand the rotational balancing.
Paul