Automation, data, and geoscience — for practitioners who build things.
Every whole-rock analysis sums to 100%. That single fact, which feels like trivial bookkeeping, breaks every Pearson correlation you ever computed on raw oxide percentages. The closure problem is the geochemist's first ghost, and it haunts modern data science wherever percentages live,...
Read more →
Most beginner Python tutorials use Iris flowers, Titanic passengers, or Boston housing prices. There is a reason a lot of geologists bounce off them, and it is not...
Read more →
Everyone tells the Abraham Wald airplane story. It's a great story. It is also slightly misleading, because the planes are obviously gone. The version of survivorship bias that...
Read more →
Closing the gap between geoscience and computer science is the work I care most about, and the coding piece is where most geologists get stuck. Tutorial hell is...
Read more →
Most exploration databases are built from samples that earned their way into the dataset. That selection process can flip correlations on their head and quietly mislead any analysis...
Read more →