I would like to know if there is a way that we can interpolate soil type from point layers?
I already have sampling data that contains coordinates, soil types, elevation.
I want to make 3d voxel layer of soil types in different height levels and trying to use Empirical Bayesian Kriging 3D to interpolate the soil type layer but it requires quantitative value only.
Please, help me if you have any idea or article that shows how to do it.
Create Thiessen Polygons (Analysis)—ArcGIS Pro | Documentation
will delineate the space based on location, then symbolize based on soil class. This will give you a first start to see what the area might look like
All of the interpolation methods in Geostatistical Analyst assume a continuous variable and will not produce meaningful results for categorical data. However, it is possible to configure Empirical Bayesian Kriging 3D to be a nearest neighbor interpolation method, where every location is given the value of its closest neighbor in 3D. This is the 3D equivalent of Create Thiessen Polygons, as @DanPatterson suggested. Calling this an "interpolation method" is a bit of a stretch, but it will produce a full 3D voxel of classified soil types, if that is what you are hoping to achieve.
A colleague recently wrote a blog outlining this workflow: