You need to have samples of the material you are going to test, this samples has to have a well know analysis by NIST or any European equivalent to NIST. Them you need to know the energy and the intensity to difract that element, the element is going to be measure and waht will be the other element that contribute to the peak. Ypu ahve to do element bu element and do a least square to find the correlation.
I've just come across this, so it may be a little late to help.
However, FYI, there are some laterite standards. If you can acquire VL-1 and VL-2 (Venezuelan laterite samples), these have been characterised by several laboratories. See
LABRECQUE, J. J. & H. SCHORIN: 1985. Some statistical parameters for
selected trace elements in VL-1 (and associated later publications).
If you are using XRF, you may have trouble with Fe-interference in the XRF response due to the high concentration of Fe in many (Indian) laterites (>>40%). if so, this has an effect on the fidelity of measurement of other transition metals (e.g., V and Cr). If so, it is better to mix your sample (pellet for traces) with 25% or even 50% pure Si, then correct the analysed values back to pure sample (100%) after analysis.
With ICP, you may have trouble with the acid digest due to presence of recalcitrant weathering residues. A sealed microwave 'bomb' can be used to take the most difficult materials into 'solution' In effect, it is more likely to be an ultra-fine suspension - so it is crucial that the prepared material is analyses very quickly after preparation (otherwise it will simply settle back out of 'suspension').
As regards the basalt protolith - I assume this may be Deccan basalt? if so, there are many published analyses of this material - it shows relatively small compositional variation. For a summary of analytical elemental ranges etc., - see my 2000 paper:
See also my papers on ResearchGate for Indian laterite and Deccan basalt analyses: (e.g., J. Pet. 2000; ZFG 2001).