An adaptable quality evaluation tool for geochemical data

Geochemical data composes compositional measurements like concentrations and isotope ratios of a wide selection of analytes. It is long tail data, with a highly complex set of analytical procedures bind each data point. This leads the data vulnerable to biases between analytical methods effects the quality of data compilations caused by these biases.

In this NFDI4Earth pilot project, we developed three tools integrated into the GeoReM database, which are made available to through the newly developed free and open GeoReM API. The three tools include two outlier tools: 1) allows users to check if a measured value for a reference material is within range of the recommended available value in GeoReM based the Horwitz outlier criteria. 2) Allows users to check if a measured value for a reference material is within the expected distribution of values in GeoReM based the interquartile outlier criteria. 3) The third tool is a data quality report tool, which dynamically calculates and returns the inter-study bias based on reference material values in GeoReM for user provided paper identifiers (DOIs) for example from GEOROC or any geochemical compilation.

The accompanying Git repository (https://git.rwth-aachen.de/nfdi4earth/pilotsincubatorlab/pilots/GEOROC) shows how to use the GeoReM API to run inter-study bias quality tools directly in a R or python work environment, presented in detail in the two demo R markdown and jupyter notebooks.