A practical, and easily accomplished, experiment to ascertain the curve would be to set up a dial gauge in a holder with a given weight attached to the plunger, lay a guitar on a surface plate so that the strings are perfectly parallel to the surface plate, (with adequate support and restraint ), and move the dial gauge along taking a reading at each fret (and intermediate points if required).

Not as accurate as a Fourier analysis, but accurate enough for all practical purposes (accurate to half a thou is good enough for me ).