Dear

You can read these papers.

If you have a finite elements mesh or any other point system where you can quantify the accelerations, you could make integrations to get displacement time data. 

From there on you could calculate the shear strains by making use of the angular distortions between the points, I guess. 

I think you can find some info in following. The latter paper is available on research gate and we sued accelerometers in the ground to look at radial dissipation of strains away from a pile

Thesis The Rapid Load Testing of Piles in Fine Grained Soils

Article Analysis of a rapid load test on an instrumented bored pile in clay

Do make sure that the accelerations are measured in a direction such that they are measuring shear strain deflections. For example, if you had a shear wave travelling in a direction x then the acceleration measurements must be accelerations in the plane of orthogonal directions y and z, and they must be separated by non zero distance in the x direction. That is, a vertically propagating shear wave cannot be assessed by two measurements in the same horizontal plane.

I wrote up a method for doing this based on vertically propagating shear waves (and accelrations measured in the horizontal plane) in physical model tests, if it helps:

Article Evaluation of Shear Modulus and Damping in Dynamic Centrifuge Tests

You can find a very good description of equivalent linear site response analysis in the manual of Strata software. I believe it will help you develop your own code.