BGS MineralsUK: UK and World Mineral Statistics

One national geological survey, two statistical families, one schema. BGS MineralsUK is the British Geological Survey's centre for economic mineral resources, and its statistical output comes in two shapes Datadory delivers together. World Mineral Statistics is the headline: an annual census of mineral production, imports and exports compiled country by country and commodity by commodity, assembled by BGS and its predecessor organisations since 1913 and machine-readable for 1970 through 2022. Precious metals sit inside it properly - gold, silver and the platinum-group metals each carry their own production tables - but the same rows also cover energy minerals, construction materials and industrial minerals, which is what makes cross-commodity comparisons possible without changing sources.

The UK Minerals Yearbook layer is the domestic counterpart: annual volumes resolving United Kingdom output into energy minerals, crushed rock, construction minerals and industrial minerals, with the latest edition released in April 2026.

Scale, concretely: millions of rows spanning tens of thousands of distinct country-commodity-year combinations, every one of them resolved against 27 documented fields. Within Datadory's catalog of 1,744 datasets across 159 viable industries, this record scores 9/10 for quality.

Get a sample of this dataset

One real record from the archive, shown flat exactly as its fields arrive:

synthetic_id      : 1-10-1995-128-2006
year              : 2006-01-01T00:00:00
table             : 128 - Mine production of silver
statistic_type    : Production
country           : Burkina Faso   (BF / BFA)
commodity         : silver, mine
erml_group        : Silver
quantity          : 0.0
book_style        : -----
precision_note    : Nil (nothing produced)
sdmx_status       : A - Normal Value
units             : kilograms (metal content)

That single row quietly settles three questions buyers always ask. First, identity: one synthetic id resolves the commodity code, the printed table and the year simultaneously, so provenance survives the trip off the page. Second, geography: the country arrives spelled out and coded to ISO 3166-1 in both alpha-2 and alpha-3, which means joins against your own tables need no fuzzy matching. Third - and this is the one that breaks naive pipelines elsewhere - absence is typed. A zero that means nothing produced carries quantity 0.0, renders as the yearbook symbol -----, says so in data_precision_description, and rides with SDMX status A, Normal Value. Missing, estimated and genuinely-nil stay distinguishable instead of collapsing into blanks.

Every other record repeats the same shape: a country with its codes, a coded commodity with an optional finer subdivision, Production or Imports or Exports, and a quantity with its unit - kilograms or tonnes of metal content depending on the commodity.

Twenty-seven documented fields, definitions verified during research - fifteen form the working core in the dictionary below, grouped into four families: identification (synthetic_id, year, the yearbook table id and its readable name), classification (statistic type, commodity translation, the European Mineral Resources Locator group), geography (country name plus ISO 3166-1 alpha-2 and alpha-3 codes), and the value complex (quantity, its book-style rendering, unit, precision description and SDMX observation status).

The remaining twelve are structural trim rather than analytical payload - geometry coordinates, deeper commodity descriptors, vocabulary links and footnote apparatus. They ship under additional fields on request rather than cluttering the default cut.

Six groups sit beyond the core dictionary and get mapped to your use case at sampling:

• Point geometry - the pole-of-inaccessibility longitude and latitude that place each country on a map, useful when the rows feed a geospatial view.
• Deeper ERML descriptors - erml_commodity and erml_sub_commodity, which spell out stage and form ("Silver (mine production, metal content)", "Ores and concentrates (metal content)") beside the bare group label.
• Vocabulary linkage - cgi_commodity_url, tying each commodity to the GeoSciML/CGI identifier that other geological datasets speak.
• Internal coding - the numeric bgs_commodity_code behind the translated name, and the optional sub-commodity split.
• Footnote apparatus - concatenated table-note and figure-note codes with their text, which explain why particular printed tables carry caveats.
• Raw status letters - sdmx_code alongside its translation, for pipelines that prefer to map status themselves.

Name the ones you want when you request the sample; they arrive as ordinary columns in the same rows, not as a side file.

Geography - worldwide on production. Every reporting country appears, from the largest producers down to economies whose entire record is one commodity line - which is precisely why a Burkina Faso silver row is a legitimate, well-formed record rather than an anomaly. United Kingdom national statistics ride alongside through the yearbook layer. Trade is wider than it later becomes: imports and exports cover all countries up to 2002, then selected European countries from 2003 through 2018.

Temporal - the deepest bench in the precious metals & minerals slice. The machine-readable layer runs 1970 through 2022; the printed World Mineral Statistics volumes reach back to 1913, making this the longest consistent cross-country production record the catalog carries. Primary aggregates begin in 1998, and the UK Minerals Yearbook editions continue through the April 2026 release.

Granularity - annual throughout: one record per country x commodity x statistic type x year. Units vary honestly by commodity and stage - kilograms of metal content for silver, different bases elsewhere - and the unit column travels with every figure, so aggregation mistakes become visible instead of silent.

API, files, or your warehouse. Daily, weekly, or hourly.

Name the commodities, the countries and the year window when you request the sample - the full cross-commodity panel, a precious-metals-only cut, or the UK yearbook layer on its own. The sample ships first either way; the ongoing feed lands on whatever cadence your models need, shaped to the scope you named rather than as an undifferentiated dump.

A century-scale mineral ledger earns its keep in five jobs:

• Market researchers and consultants - the citable base table for mineral market studies. Production, trade and country coverage measured one way by one survey is the exhibit reviewers stop arguing about.
• Investors and quant researchers - the supply side under the price. Country-level gold, silver and PGM production panels give metals models exogenous factors with enough history to backtest against.
• Journalists, academics and students - provenance that traces to a national geological survey and, before 2023's machine-readable era, to printed volumes still on library shelves.
• Competitive intelligence teams - production and trade shifts by country evidence supply concentration and jurisdiction risk in client briefs with numbers rather than adjectives.
• Data scientists and ML engineers - typed rows with ISO-coded geography and explicit nil semantics, so feature engineering starts at modelling rather than at cleaning.

Deeper guidance sits on the market researchers, investors & quants and journalists & academics persona pages.

Market Researchers & Consultants get their three qualifying questions answered from the coverage alone - scope, geography and years available in one place, at maximum relevance for sizing work. Journalists, Academics & Students get the original rather than someone's chart of it, and the attribution that survives peer review. Investors & Quant Researchers get a sample period long enough to matter: five decades machine-readable, a further six in print. Data Scientists & ML Engineers get typed rows where nils declare themselves, so a zero never masquerades as data loss mid-pipeline. Competitive Intelligence & Product Teams get change detection across whole countries rather than single companies. Start from the precious metals & minerals data hub, then read vs USGS Mineral Commodity Summaries for where the long archive beats the rolling five-year window.

Provenance note - compiled by the British Geological Survey, the national geological survey of the United Kingdom, continuing a World Mineral Statistics series its predecessor organisations began in 1913. One editorial hand across the whole run is why the country-commodity-year shape holds for decades at a stretch.

Methodology note - figures are compiled from returns and published sources into printed yearbook tables, and the fields preserve that heritage: book-style symbols for absent figures, precision descriptions for special cases, SDMX status codes for observation quality. Read quantity together with its precision description rather than alone. Trade narrows after 2002 to selected European countries through 2018, and primary aggregates start in 1998.

Completeness note - Datadory scores this record 9/10, and the fifteen-field core above maps fully to the verified dictionary. Availability of machine-readable years after 2022 is confirmed at sampling rather than asserted here, and pre-1970 depth arrives from the printed archive.

Where to go next - pair the long archive with a rolling-window rival and the wider trade lens: USGS Mineral Commodity Summaries for the current-year read, UNCTADstat and OEC for trade values, the Indian Bureau of Mines for district-level depth in one large market. The glossary entries below define the vocabulary the tables use.