World Gold Council GoldHub Data Hub: Fourteen Series, One Gold Ledger

One trade association keeping its market's entire ledger, fourteen series deep. The World Gold Council publishes Goldhub, its data hub, as fourteen named series organised into four groups: price and premiums, performance metrics, demand and supply, and central banks.

The price group runs from spot-price averages - monthly, quarterly and annual, in major trading, producer and consumer currencies, back to 1978 - through local premium and discount readings comparing Indian and Chinese consumer prices with the international dollar price, to futures prices across key exchanges. Performance metrics compute monthly returns for gold against bonds, equities and other assets, annualised volatility, correlations to other asset classes and aggregated trading volumes.

The demand-and-supply group is the heart of the hub: comprehensive demand by sector and country balanced against mine production, recycling and producer hedging, most recently refreshed 30 July 2026 with Q2 2026 tables - plus ETF holdings and flows, mine production by country, all-in sustaining production costs with a latest-quarter cost curve, futures positioning across nine major exchanges, and above-ground stock tallied since 2010. The central-bank group closes it out: official reserves by country with attribution of sales under the Central Bank Gold Agreement, and the annual Central Bank Gold Reserves Survey, run since 2018.

Scale, concretely: fourteen series whose individual workbooks reach up to roughly forty-eight years of observations, every observation typed against five 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

Two row shapes carry the hub, shown flat exactly as their fields arrive:

# One quarter off the demand-and-supply balance, flat, exactly as the fields arrive
date            : 2026-Q2
demand_sector   : Jewellery
country         : China
tonnes          : <quarterly tonnage for the sector-country cell>

# One month off the spot-price workbook - same spine, different measure
date            : 2026-07     (period identifier; grids run monthly | quarterly | annual)
currency        : USD         (columns run USD | EUR | GBP | JPY | INR | CNY ...)
price           : <average dollars per troy ounce>

# The shape every other row repeats
period          : <monthly-quarterly-annual price grids | quarterly balances | monthly ETF flows and reserves snapshots>
dimension       : <demand sector | country>, wherever the series splits that far
measure         : tonnes for demand, supply, holdings, stock and reserves; local-currency price per troy ounce for price rows

Three buyer questions settle themselves here. First, spine: whatever the series - price, balance, holdings or reserves - every row keys on a period identifier, with sector and country dimensions joining wherever the series splits that far. Second, measure discipline: quantities travel in tonnes and prices per troy ounce in whichever currency column you read, so a 224.5 never gets mistaken for a price and a 3,300 never gets summed onto a weight. Third, clock: monthly price grids sit above quarterly balances and monthly ETF-flow snapshots, so aggregation direction is explicit rather than guessed.

Every other series repeats this anatomy - futures positioning, cost curves and the reserves tables simply swap the dimension set while keeping the same spine.

Five documented fields form the shared spine, definitions inferred during research from the series descriptions. Date carries the period identifier on monthly, quarterly or annual grids. The currency price columns hold average gold prices per troy ounce in each published currency - dollars, euros, sterling, yen, rupees, yuan and their neighbours. Demand sector classifies consumption as jewellery, technology, investment in bars, coins and ETFs, and central banks together with other institutions. Country drives the demand tables, mine production and the reserves series. Tonnes measures everything counted rather than priced.

Those five are the intersection, not the entirety. Each of the fourteen series layers its own columns over this spine - futures positioning adds exchange-level open interest across nine venues, production costs add all-in sustaining cost lines and the latest-quarter cost curve, the above-ground-stock series adds its post-2010 breakdown - and those column families are mapped to your use case at sampling rather than asserted here.

Geography - global first, national where it matters. Price series quote worldwide in multiple currencies, while the demand tables, mine production and central-bank reserves break out by country, so both a dollar-denominated world view and a named-country argument read off the same hub.

Temporal - the long game. Prices run from 1978 onward, above-ground stock is tallied from 2010, the quarterly supply-and-demand balance reaches back across decades, and the central-bank survey has run annually since 2018. One horizon worth planning around: historical LBMA Gold Price data was withdrawn at ICE Benchmark Administration's request, so from 18 March 2025 only limited LBMA price data remains inside the hub.

Granularity - three clocks in one place: daily-to-annual price averages, a quarterly supply-and-demand balance, and monthly ETF-flow and reserves snapshots. Aggregating upward is arithmetic; disaggregating downward is impossible, so name the finest clock your model needs when requesting the sample.

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

Name the series, the dimensions and the span when you request the sample - the full quarterly balance by sector and country, the reserves table with survey results attached, or just the multi-currency price grid across your lookback. 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 ledger that both prices and counts the same metal earns its keep in five jobs:

• Central-bank behaviour analysis - reserves by country with Central Bank Gold Agreement attribution, paired with the annual survey of reserve managers since 2018, turns accumulation headlines into dated tonnage and stated intent.
• Balance modelling - jewellery, technology, investment and institutional demand against mine production, recycling and producer hedging, quarterly and by country, is the raw material of deficit and surplus arguments.
• Multi-currency price research - averages back to 1978 in trading, producer and consumer currencies, with Indian and Chinese local premia beside them, separate a gold story from a currency story.
• Miner economics - all-in sustaining costs and the latest-quarter cost curve beside production by country put margins, not just tonnage, on the record.
• Portfolio evidence - monthly returns, volatility and correlations against bonds and equities give allocation models a gold leg built on the market's own numbers.

Deeper guidance sits on the investors & quants, market researchers and data scientists persona pages.

Investors & Quant Researchers get the factor-ready layer - returns, volatility, correlations and a decades-deep price history - with the fundamental ledger underneath. Market Researchers & Consultants get demand by sector and country cited to the body that compiled it, which is provenance peer scrutiny accepts. Data Scientists & ML Engineers get a five-field spine repeated across fourteen series, so joins become configuration. Competitive Intelligence & Product Teams get authoritative context for anything gold-adjacent, from jewellery sizing to vault products. Developers & Builders get flat period-keyed rows that drop straight into dashboards. Start from the best precious metals & minerals datasets shortlist, then read vs BullionVault gold prices for where the structural ledger beats the intraday tape.

Provenance note - published by the World Gold Council through its Goldhub data hub, with market data credited to ICE Data Services. An industry body compiling its own market's fundamentals is first-party provenance, and the embedded market-data credit belongs in any citation.

Methodology note - these are period statistics, not tick captures: the price grids publish monthly, quarterly and annual averages, and the balance is quarterly. Definitions for the five core fields were inferred from series descriptions rather than read off workbook internals, so confirm exact column semantics at sampling before hard-coding a parser. Mind the seam noted under coverage, too: LBMA price history inside the hub thins after 18 March 2025.

Completeness note - Datadory scores this record 9/10 against a catalog mean of 7.81 across 1,744 datasets, among the top-scored records in the 27-record precious metals & minerals slice. Breadth is the asset: no other single record here puts prices, balances, costs, flows and reserves in one place. If your question needs intraday shape or session-level prints, the cards below carry that lane.