Where uncommon minerals should be, and what the old surveys described without having a word for it.
Emblem XXI, the squared circle. Michael Maier, Atalanta Fugiens, 1617.
inerals are unevenly distributed in a way that is partly geology and partly attention. Some localities produce a long tail of species found nowhere else, and the mineral-ecology literature gives you enough structure to ask where the next ones should be rather than waiting to be told.
The second thread is archival. Victorian and early twentieth century geologists wrote down, in prose, a great many observations that no structured database ever captured, because the categories did not exist yet. Troublesome black sand in a tailings pile is a sentence in a bulletin and a rare earth assay today.
The Work
Single-Locality Minerals
Active
Predicting where more single-locality mineral species should be found.
Classifies the top hundred type localities into two dozen geological archetypes, models endemic species counts as type-locality count multiplied by an archetype retention rate, and ranks global sites predicted to yield more species found nowhere else. Grounded in the Hazen mineral-ecology work rather than invented from scratch.
What existsSmall and complete as far as it goes: two scripts, a built database, and a written report producing 56 ranked hotspots. The report documents its own principal weakness, which is that the best source data sat behind login walls and refusals, so the input is thinner than the method deserves. No visualization exists yet.
Critical Minerals in the Old Bulletins
Active
Reading century-old USGS reports for byproduct minerals nobody was looking for at the time.
The engine reads nineteenth and early twentieth century bulletins and annual reports for narrative descriptions of material that was a nuisance then and is a critical mineral now, then translates that prose into modern chemistry and ranks abandoned workings worth re-examining. The interesting claim is not about any particular mine. It is that a structured-database view of mineral occurrence silently dropped everything that was only ever described in sentences.
What existsA packaged pipeline with 17 commits, last touched in April. The intended deliverable was a defensible ranked list of abandoned workings, and I did not finish it. Treat this as a method that runs rather than a result.
inerals are unevenly distributed in a way that is partly geology and partly attention. Some localities produce a long tail of species found nowhere else, and the mineral-ecology literature gives you enough structure to ask where the next ones should be rather than waiting to be told.
