Supporting materials for classification models for the total sounding. Source data from NPRA's GUDB database.

Raw data from soil investigations copied from GUDB in the summer 2020.  The dataset has been scrubbed to only contain data from field- and laboratory tests, removing reports, text files and other data types that might contain personal information.

The dataset contains 250331 files in 3284 folders (with extensions .snd, .tlk, .prv, .gvr, .cpt, .tot, .std, .dpt, .log, .kof, .rko, .rla, .dtr, .rdt, .rdp, .pvt, .pvs, .vct, .org, .vim, .xml,). The files are mostly stored in a human readable state, meaning this 760MB zip unpacks to about 4.6GB on disc.

Please see the attached dataset components figure to see the distribution of the different file types.