Steel Statistical Yearbook Data

It is the deepest shelf in the steel pool, delivered here as typed rows instead of a statistical volume you mine by hand. The Steel Statistical Yearbook is how the World Steel Association publishes the statistics exchanged with its member associations: 59 standard tables covering roughly ninety countries and economies across 2015-2024, with regional totals built from the listed countries only.

The table map is the value. Production runs from total crude steel through by-process splits (BOF, electric furnace, other), ingots, continuously cast output and liquid steel for castings, with one table carrying monthly crude steel production. Hot-rolled products divide into long (railway track material, heavy and light sections, concrete reinforcing bar, wire rod), flat (plate, coil/sheet/strip, electrical sheet, tinmill, coated sheet) and tubular (seamless, welded). Trade tables pair imports and exports for semis, long, flat and tubular products, ingots/semis and pig iron. Use tables state apparent steel use in crude-steel-equivalent and finished-product terms, per-capita included, then push further into indirect trade and true steel use. A raw-materials block closes the loop: pig iron, direct reduced iron, iron ore and scrap, each in production and trade.

Scale, for perspective: world crude output ran near 1,849 million tonnes in 2025 (provisional), with worldsteel tallies of apparent steel use near 1,885 million tonnes in recent years - the demand side of the largest manufactured-materials market on earth lives inside this spine. Within Datadory's catalog of 1,744 datasets across 159 viable industries, this slice scores 7/10 - the widest indicator set in the pool, one authoritative issuer, and notation conventions stated inside the publication itself.

Exactly as they land after normalization - the spine every table shares, the notation conventions that govern every figure, and the 59 tables grouped as they run:

# the spine every table shares
row_key       = country_or_area x year x measure (production | import | export | use)
unit_default  = thousand metric tons, unless a table states otherwise
regional_note = totals comprise listed countries only

# notation conventions printed inside the edition
value 0       -> less than 500 tonnes
value "..."   -> not available
suffix e      -> worldsteel estimate for that figure
(e) after name-> whole series estimated for that country

# the 59 tables, grouped as they run
production     total crude | by process: BOF / electric furnace / other
               ingots | continuously cast | liquid steel for castings
               monthly crude steel production
long_products  railway track material | heavy & light sections | rebar | wire rod
flat_products  plate | coil/sheet/strip | electrical sheet | tinmill | coated sheet
tube           seamless | welded
trade          imports & exports: semis, long, flat, tubular, ingots/semis, pig iron
use            apparent steel use: crude-steel-equivalent & finished terms, per-capita
               indirect trade | true steel use
raw_materials  pig iron | DRI | iron ore | scrap - production and trade
annex          sources and definitions

# shape of the slice
scale=~90 economies x ~60 indicators x 10 years   span=2015-2024

Read what the notation block already buys you. A zero in this compendium is not a missing value - it certifies output below 500 tonnes, which is the difference between "no industry" and "a niche one" in forty smaller economies. An estimate flag tells you whose number you are holding, the reporter's or the association's. Because those conventions are part of the source itself, deliveries preserve them per cell instead of flattening decades of statistical practice into a bare numeric column. Request a sample naming your economies and product families and the same shapes come back filled.

Four spine columns verified against the edition, plus the structural dimensions the tables hang their measures on. Every observation keys on country or area - a country, an economy grouping or a regional aggregate whose totals comprise listed countries only - paired with the year of observation, 2015-2024 in the current edition. The measured value arrives in quantity, thousand metric tons unless a table states otherwise, and carries its estimate flag: a suffix marking a figure the association itself filled in, or a marker showing a whole national series is estimated.

Around that spine sit the dimensions that make the 59 tables addressable separately: process route where production splits by converter, electric furnace and other routes; product family where output divides into long, flat and tubular lines; measure type distinguishing production from imports, exports, apparent use and true use. Per-capita derivatives ride beside their absolute series in the use tables rather than requiring a population join on your side.

The dictionary below is the verified core. Anything adjacent confirmed during sample preparation - finer product detail inside a family, additional economy groupings - folds under additional fields on request rather than being promised blind.

Geography: roughly ninety countries and economies worldwide, the widest single-issuer country surface in the steel pool, plus regional aggregate rows. The aggregate rule matters: totals comprise listed countries only, so a region subtotal never silently absorbs a non-reporting economy - you always know exactly what a total sums.

Temporal: 2015-2024 in the current edition - a clean ten-year window per country and indicator, deep enough for cycle work and structural comparison alike. Earlier editions extend the reach further back, so longer histories are a matter of stacking vintages; the exact earliest span of any given table gets quoted at sample scoping rather than guessed here.

Granularity: annual by country and product throughout, which is the grain strategy decks and market models actually want, stepping down beneath the headline to process route and product family, and sideways into trade cells by product group. The one deliberate break in rhythm is the monthly crude steel production table - month-by-month texture inside an annual compendium, handy for checking whether an annual move was steady or concentrated in a quarter.

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

Pick the channel your stack already speaks. The rows arrive identical either way - economy, year, measure, tonnage, estimate flag - with the tabular source material already extracted and typed, so nothing in your pipeline re-keys figures out of a statistical volume.

Every delivery ships with the field dictionary above plus the sample rows for validation, and the notation travels on the data itself: estimate flags stay attached to the cells they cover, and the distinction between a reported zero and an unavailable cell survives into your warehouse.

• Market researchers and consultants treat it as the reference volume for steel-intensive markets - apparent steel use in finished-product and per-capita terms sizes demand from the consuming side; worked flows continue on our market sizing page.
• Trade-policy and competitive-intel analysts read the import-export matrices by product family, where a flat-products deficit tells a different story than a semis one; positioning work pairs it with competitor tracking.
• Investors and quant researchers build country-year panels across ninety economies with process-route splits attached - the converter-versus-furnace mix separates integrated from scrap-fed supply; forecasting workflows live on demand forecasting.
• Supply-chain and procurement teams map upstream exposure beside the mills: pig iron, direct reduced iron, iron ore and scrap in one spine with the steel they feed; see supply-chain mapping.
• Journalists, academics and students cite the industry's own statistical authority with per-capita and true-use measures ready-made; provenance-first workflows continue on citation-grade research.

Market researchers and consultants get the global reference volume - ninety economies against sixty indicators in one issuer-consistent frame, the base layer any steel market model stands on; the fuller workflow lives on market researchers in steel. Investors and quants get country-level depth the monthly pulse cannot carry - process route, product family and use measures stacked ten years deep; see investors and quants in steel. Data scientists and ML engineers get a compact, fully numeric macro feature whose estimate flags make train/serve hygiene explicit and whose joins need only economy and year; see data scientists in steel. Journalists, academics and students get citable official statistics with the association's own definitions annex standing behind every number.

Provenance note — compiled by the World Steel Association (worldsteel) from statistics exchanged with its member associations, and published with an annex of sources and definitions inside the volume itself. One issuer stands behind all 59 tables, which is precisely what makes cross-country comparison safe.

Methodology note — the notation is the methodology contract: thousand metric tons unless stated otherwise, zero meaning under 500 tonnes, "..." meaning unavailable, a suffix marking association estimates, and a parenthetical marker flagging whole estimated series. Regional totals comprise listed countries only. Deliveries preserve every flag rather than silently overwriting history.

Completeness note — Datadory scores this record 7/10 and the field dictionary above verifies against the current edition reviewed during cataloging. Archive depth before the current ten-year window and any finer product detail inside a family are confirmed at sampling rather than asserted here.

Where to go next — pair the deep shelf with the frames around it: monthly crude steel production for 71 countries is the running pulse at roughly 98 percent of world output; UNSD Industrial Commodity Statistics (steel) widens the crude-series country count past a hundred; iron and steel statistics and information holds the century-long American spine; best steel datasets ranks the whole vertical; and the steel data hub pools the industry view.