THE APPROACH / STATISTICAL GEOMETALLURGY
Statistical thinking.
Practical decisions.
At Clauson Geomet, statistical geometallurgy means applying statistical thinking to the whole journey from geological and processing data to a decision: how data are sampled, structured, checked, analysed, interpreted and put to work.
Better data. Not just better models.
Statistics does not start with fitting a model, or end with reporting one. I bring the same care to data organisation, quality control, visualisation and communication as to the analysis itself — building reproducible workflows that others can understand, check and reuse.
The whole-workflow perspective draws inspiration from Hadley Wickham and the tidyverse community: data organisation, exploration, programming and communication are integral to analysis, not just preparation for a model.
Small samples. Careful inference.
Geometallurgical decisions often rely on limited testwork. Drawing conclusions beyond those samples requires an understanding of the orebody, the sampling and the process, alongside appropriate statistical techniques.
Experience guides the questions and assumptions; it does not replace representative sampling, validation or honest uncertainty. A model can fit the data and still answer the wrong question. The difference is knowing what the samples represent, how the process works and which conclusions the evidence can support.
Rigorous methods. Useful tools.
I turn statistical approaches into practical, problem-specific implementations: reproducible R and Python workflows, interactive applications and decision-support tools. Where standard approaches fall short, I develop and test new implementations around the problem — with clear assumptions, documented methods and outputs people can use.
The same perspective underpins my applied research in compositional data analysis and causal inference.