EU Business Demography — Formation & Survival, Per Record

Turn Eurostat's business-demography cube into clean, per-record company-formation and survival benchmarks — one row per published observation, broken down by business-demography indicator, enterprise size class, industry (NACE Rev.2) and country, ready to size company formation, death and survival across Europe.

Each row of the source multidimensional cube becomes one clean, flat record: every dimension is expanded into an explicit code and its label, the observation value is reproduced exactly as Eurostat published it, and each record is stamped with source, dataset code, licence, attribution, a UTC retrieval timestamp and an interpretation caveat.

What one record represents

The source dataset is Eurostat bd_9bd_sz_cl_r2 — business demography by size class (NACE Rev.2): enterprise births, deaths, survival and active populations. Eurostat publishes it as a JSON-stat hypercube indexed by several dimensions (frequency, business-demography indicator, size class, NACE Rev.2 activity, geography and time); this Actor unpacks it so every present observation cell becomes exactly one record. Cells Eurostat does not publish are absent from the source and are neither invented nor delivered; published zeros are kept as real values.

For each record you get a composite record_id, the dataset_code, the indicator name, then — for each dimension — the code and its label, plus time_period, the numeric value, and Eurostat's obs_status flag when present. The dimension columns include indic_sb (the business-demography indicator — active enterprises, births, deaths, survival counts and rates) with its label, sizeclas (enterprise size class), nace_r2 (industry) and freq.

Coverage and volume

Per licence-permitted territory the cube is large: Germany alone carries 501,603 non-null observations across the published years. Across all licence-permitted territories the full-cube capacity is 15,695,464 non-null observations spanning 2004-2020.

Live full-cube allowed count is 15,695,464 records across 2004-2020 and all licence-permitted territories; the Wave-3 index figure of 1,234,612 reflects the 2020 single-year slice only. This README states the difference; the capacity line is the full-cube count.

Because the full multi-territory cube exceeds the dissemination API's single-response size limit, this Actor takes the reporting country as a required input (prefilled to DE), pulling one territory at a time; the generalized ingester also auto-partitions on a size limit where needed and stops at your Maximum records cap.

Licence and attribution

This dataset is redistributed under the terms Eurostat publishes it with; the full notice travels on every record in the licence field:

Eurostat, CC BY 4.0. Reformatted from Eurostat JSON-stat; values unmodified. Third-party content and logos excluded. Commercial-barred geographies removed.

A structural geography allow-list is applied at ingest: only EU member states, EU/euro-area aggregates, EFTA members and official EU acceding/candidate countries are delivered; commercial-barred territories are excluded at ingest and counted in the run receipt. The required attribution — Source: Eurostat (dataset bd_9bd_sz_cl_r2) — is on every record.

Interpretation caveat

Business demography indicators (enterprise births, deaths, survival, active-enterprise populations and related rates) by size class and NACE Rev.2 activity. Counts are enterprise counts, not turnover or employment; read the indic_sb code to know which measure a row carries.

Eurostat values are reproduced verbatim: no rescaling, re-basing, interpolation or seasonal adjustment. Read each value together with its dimension labels — the indic_sb code tells you which measure a row carries.

Person-data policy

This is a statistics product about enterprises, industries and territories. There are no natural-person or contact fields anywhere in the record, and a per-record assertion enforces that allow-list at write time.

Provenance, licensing and compliance

Every run begins with a live source-preflight: the Actor reads the exact host's robots.txt at runtime and refuses to proceed if the crawl policy disallows the data path. The gate result — URL, HTTP status, byte length and a SHA-256 of the policy — is written to the run's RUN_RECEIPT, so each run carries its own audit trail. The Actor identifies itself with a transparent, non-impersonating User-Agent and never bypasses a block, solves a challenge, or fetches through a cache or mirror. When the door is genuinely unavailable the run fails loudly and bills nothing.

Inputs

• Country / territory (geo) — required; the Eurostat geo code to pull (e.g. DE). This cube is too large to return in one request, so it is retrieved one territory at a time.
• Earliest period (sinceTimePeriod) — return observations from this year onward. An earlier start yields more records.
• Maximum records (maxRecords) — hard cap on records delivered and billed.

Output

Records land in the Actor's default dataset and export as JSON, CSV, Excel or via the Apify API. A tabular overview view surfaces geography, period, the key dimension labels and the value.

Fields in detail

Each record leads with a composite record_id built from every dimension code, so it is stable and join-ready across runs. Every dimension arrives as both a machine code and a human-readable label, which means you can group and filter on the codes while reading the labels — no separate code list to maintain. The value is the observation itself, reproduced exactly; obs_status carries Eurostat's flag (for example a break in series or a provisional marker) when one is present, so you never silently lose the caveats Eurostat attaches to a number. The provenance block — source, source_dataset, licence, attribution, caveat, observed_at — is present on every record, so each row is self-describing and audit-ready without reference back to this page.

Sibling Actors

This Actor sits alongside the fleet's national-macro-indicator-records cell — but that carries macro aggregates (headline national indicators), while this one is enterprise-level business demography (formation, death and survival by size and industry): a different grain and a different question. It also shares its engineering — the runtime robots gate, the geo allow-list, the push-then-charge billing and the verbatim-value discipline — with the fleet's other Eurostat cells.

Pricing

This Actor uses Apify's pay-per-event model: a flat $0.05 per record actually delivered to the dataset, and nothing else — no monthly rental, no per-run base fee, no compute charge. Deliver 40 records and you pay $2.00; deliver 10,000 and you pay $500.00. Billing is wired after delivery — each record is pushed first and only then does the per-record event fire — so a mid-run failure can only ever under-charge you, never over-charge. Use Maximum records to cap spend precisely.

Scaling and limits

Set Maximum records low to sample cheaply or high to pull a territory in full. The Actor streams the cube and delivers incrementally, so memory stays bounded regardless of how large the requested territory is, and you are billed only for records actually delivered. Because Eurostat updates this cube on its own release calendar, re-running the Actor picks up new periods automatically, and each record's observed_at stamp dates the snapshot — schedule a run to keep a downstream table current. To assemble a multi-country panel, run the Actor once per territory and concatenate; every record already carries its geo code and label, so the panels align without any reconciliation work.

Typical uses

Size company formation and closure by country and industry; measure survival rates by cohort; compare small-firm versus large-firm dynamics; benchmark entrepreneurial activity across member states; or feed a strategy, credit-risk or market-entry model with clean, licence-clean business-demography inputs.

What this Actor does not do

It does not forecast, model, seasonally adjust, or merge multiple Eurostat datasets into one record, and it does not deliver commercial-barred geographies. It gives you faithful, licence-clean, analysis-ready observation records — one row per published cell — with an auditable provenance trail.