Wikidata Mineral Coverage Is Oxygen-Heavy and Lanthanide-Sparse

Ledger + structural data-quality observation. The ledger is the per-element mineral count for the 72 elements that appear in any Wikidata mineral formula. The headline is the dominance of oxygen — 4,588 of 5,662 formulas (81.0 %) contain it — followed by hydrogen (55.4 %), silicon (28.2 %), calcium (24.7 %), iron, sulfur, aluminum, sodium, magnesium, copper. The average Wikidata mineral is essentially a hydrated silicate. The structural observation is the 46-element absence list: noble gases (He, Ne, Ar, Kr, Xe, Rn — chemically inert), naturally short-lived elements (Tc, Pm, Po, At, Fr, Ra, Ac, Pa, Np, Pu, Am, Cm, ...), all transactinides (Rf through Og), and six naturally-occurring lanthanides — europium, terbium, holmium, erbium, thulium, lutetium. The lanthanide absences are physically wrong: these elements do exist in real mineral deposits (the rare-earth ore monazite, for example, contains all of them in trace amounts). They are missing from Wikidata's mineral formulas because mineralogical convention writes rare-earth content as 'REE' or '(REE)' in formulas rather than enumerating individual lanthanides. So the absent-lanthanide list is a clean fingerprint of the catalog's curation rule rather than a fact about Earth's chemistry. Pinning the specific 81 % oxygen prevalence and the specific 6-lanthanide hole gives mineralogists and data curators a snapshot-pinned reference for the Wikidata mineral catalog's coverage as of 2026-04-13.

Earth has about 6,000 different known minerals — silicates, oxides, sulfides, salts, and so on. Each mineral has a chemical formula like 'CaCO₃' (calcite) or 'SiO₂' (quartz). I downloaded the chemical formulas for every mineral species in Wikidata (over 5,600 of them) and asked: which chemical elements are in the most minerals? The answer is overwhelming. Oxygen is in 81 % of all formulas — over four out of every five minerals on Earth contain oxygen. Then hydrogen at 55 %, silicon at 28 %, calcium at 25 %, and iron, sulfur, aluminum, sodium, magnesium, copper after that. So the 'average' mineral on Earth is basically a silicate that contains some water. None of that is surprising to a geologist. The surprising part is the absence list. Of the 118 elements on the periodic table, only 72 appear in any Wikidata mineral formula. The other 46 are absent. Some absences are obvious — the noble gases like helium, neon, argon don't bond chemically and almost never form minerals. The transuranic elements (above uranium on the table) are too radioactive and short-lived to exist in nature. But six elements I expected to find are missing too: europium, terbium, holmium, erbium, thulium, and lutetium. These are 'rare-earth elements,' and they DO exist in real ore deposits — the monazite mineral, for instance, contains all of them in small amounts. They're missing from the database because mineralogists conventionally write 'REE' (rare-earth element) in mineral formulas instead of writing out each individual lanthanide. So when I count which elements appear in mineral formulas, those six lanthanides come up as zero — not because they don't exist in real minerals, but because the catalog's convention erases them. Two findings in one: oxygen is everywhere, and Wikidata has a fingerprint of the way mineralogists abbreviate rare-earth content.

4588 O · 3135 H · 1595 Si · 1399 Ca · 1248 Fe · 1209 S · 1158 Al · 1102 Na · 791 Mg · 748 Cu

5,662 mineral formulas · 72 of 118 elements present · 46 elements absent

Eu (europium), Tb (terbium), Ho (holmium), Er (erbium), Tm (thulium), Lu (lutetium) — present in real REE ores but masked by 'REE' shorthand in catalogue entries