EU steel & basic-metals production statistics (dissemination API)

It is the European statistical system's account of basic-metal manufacturing — NACE Rev. 2 division C24, the division holding iron and steel and ferro-alloys at C24.1. Its spine is the short-term statistics family sts_inpr_m and sts_inpr_a, which carry the PRD production-volume indicator for industry: a dimension-coded index with base years including 2021=100, seasonal-adjustment flags separating non-adjusted, calendar-adjusted and seasonally-and-calendar-adjusted readings, and history reaching back to 1953 for some geographies.

Around that spine sit three complements. Annual detailed enterprise statistics (sbs_na_ind_r2) add hundreds of structural indicators — persons employed, turnover, value added, wages bill — down to C24 subclasses across NACE B–E. Employment by detailed industry arrives through nama_10_a64_e. PRODCOM carries product-level output, and the Comext trade family tracks EU steel flows with partner detail.

Datadory scores this record 9/10 inside a catalog of 1,744 datasets across 159 viable industries — one of twenty-three steel records. Where national federations summarize, this is the underlying official measurement, dimension-coded and joinable.

Get a sample of this dataset

Three monthly index observations for German basic metals and one structural-layer row — values as the dimensions encode them:

dataset : sts_inpr_m      freq : M    indic_bt : PRD
nace_r2 : C24             s_adj : NSA unit : I21 (2021=100)
geo     : DE              time : 2026-01   value : 88.1
geo     : DE              time : 2026-03   value : 97.6
geo     : DE              time : 2026-06   value : 92.9

dataset : sbs_na_ind_r2
label   : Annual detailed enterprise statistics for industry (NACE Rev. 2, B-E)
nace_dimension_size : 414
updated : 2024-03-01T23:00:00+0100

Read the first family left to right and the mechanics are visible: an unadjusted (NSA) German index on the 2021 base dips to 88.1 in January, rebounds past its base-year level to 97.6 by March, then settles at 92.9 in June — precisely the shape that makes the seasonal-adjustment flag worth filtering on before anyone calls a turning point. The second family is the structural layer: 414 NACE classes wide, annual, with person-, turnover- and value-added-level indicators behind every cell. Request a sample naming your geography and adjustment preference and the same two shapes come back filled for your markets.

Nine fields carry every observation, and they behave like a coordinate system rather than a spreadsheet: choose a frequency, an indicator, a classification slice, an adjustment state, a unit and a geography, and each reference period resolves to exactly one value. The dictionary below maps all nine.

Because the fields are shared across the whole short-term family, a German C24 row and an EU27 aggregate row differ only in their coordinates — no re-mapping between them. Derived views sit under additional fields on request: PRODCOM product-level output mapped to the active C24 product codes, net-trade derivations from the Comext flows, and index-versus-turnover joins against the structural layer. They are named at sampling and computed on delivery rather than approximated here.

Geography — EU27_2020 aggregate, euro-area aggregates, EFTA members and candidate countries. Country-level depth varies by indicator: 23 geographies carried C24 monthly data in the most recent months tested, so a union-wide read needs the aggregate checked against its members.

Temporal — monthly series reach from January 1953 to the current month, one of the longest machine-readable production archives in the catalog. The structural business-statistics panel covers 2005 through 2020.

Granularity — monthly on the short-term side, annual on the structural side; NACE Rev. 2 detail descends to C24 subclasses, and the structural layer adds firm-aggregate economics beneath the volume index.

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

Name the geographies, adjustment states and base year when you request the sample — a single-country index or the full C24 estate across the union, monthly or annual or both. The sample ships first either way; the ongoing feed lands on whatever cadence your models need, in whatever form your stack already reads.

An official production index earns its keep in four specific jobs:

• Market researchers and consultants — the chapter-one exhibit for any European steel study. A union-wide volume index on a fixed base, decomposable to member states, is the trend line everything else in the deck hangs from.
• Investors and quant researchers — cycle timing with a control built in. Because the non-adjusted and adjusted readings travel as separate dimension values, you can test whether an apparent turning point survives seasonal correction instead of inheriting someone else's smoothing choice.
• Economists and policy teams — the structural layer connects the volume story to persons employed, turnover, value added and wages bill for the same C24 classes, so a production drop can be traced to its labour and revenue consequences in one dataset family.
• Data scientists and engineers — a dimension-coded archive back to 1953 with stable codes means feature pipelines join on coordinates, not fuzzy entity resolution, and backtests get seventy years of regime variety to train on.

Market Researchers & Consultants get scope answered from the record itself: geography list, temporal range and granularity in one place, with the citation weight of official European statistics behind every figure. Investors & Quant Researchers get a sample period long enough to matter — seven decades of index history is enough runway to test whether production momentum actually led prices. Data Scientists & ML Engineers get cleanly coded dimensions and two complementary layers, high-frequency signal plus slow-moving structure, from one record. Journalists, Academics & Students get the original measurement rather than a republisher's chart of it, which is what review asks for. Start from the steel data hub, then read vs Steel Import Monitor (SIMA) for where European production measurement beats a US import lens.

Provenance note — published by Eurostat, the statistical office of the European Union, whose databrowser indexes the industry theme these datasets belong to. One classification standard, NACE Rev. 2, runs across the short-term, structural, PRODCOM and trade layers.

Methodology note — the production index is a volume measure on a fixed base (currently 2021=100, with earlier bases retained as separate units). Seasonal adjustment comes in three states, and the unadjusted variant is the one that preserves the raw winter dip visible in the sample rows. Percentage-change measures ride along as units, not derivations you must compute.

Completeness note — Datadory scores this record 9/10 and the field dictionary above maps fully to the published dimensions. Two edges stay honest: the active PRODCOM product codes for C24 sold quantities are confirmed at sampling rather than guessed, and EU27_2020 aggregate values for recent months are verified there too, since individual countries sometimes publish ahead of the union aggregate.

Where to go next — pair the official index with industry-side and global views: EUROFER's European steel data for the producers' own account, worldsteel's monthly crude-steel series for 71 countries, UNSD's industrial commodity statistics for a UN-frame comparison, and USGS iron and steel statistics for the US counterpart. The glossary entries on the industrial production index, short-term business statistics and NACE divisions explain the machinery underneath.