Eurostat Scraper: EU Statistics & Economic Series

Eurostat Scraper extracts official European statistics, GDP, employment, trade, energy, demographics, prices, straight from the Eurostat dissemination API, and turns the raw JSON-stat 2.0 hypercube into flat, labelled rows you can put in a spreadsheet.

Eurostat does not publish rows. It publishes a cube: a flat map of numbers keyed by a single linear index, plus a separate block describing the axes. Nothing in the payload pairs a number with the country and year it belongs to. This Actor does that reconstruction for you, correctly, and hands back one record per observation with every dimension resolved to both its code and its human-readable label.

Agent skill: SKILL.md

https://api.apify.com/v2/key-value-stores/t7YoTxpZEJOWvw4Ug/records/eurostat-scraper.md

What you get

Output field	Meaning
dataset	Eurostat dataset code, e.g. nama_10_gdp
datasetLabel	Full dataset title in your chosen language
geo, geoLabel	Territory code and its name, DE / Germany, EU27_2020 / European Union - 27 countries (from 2020)
time	Period as Eurostat writes it: 2023, 2023-Q1, 2023-01
value	The observation itself, as a number
unit	Unit of measure label, e.g. Current prices, million euro
dimensions	Every dimension of the cube, each with code, label and dimensionLabel, nothing from the source is dropped
flags	Observation status marks, decoded, p → provisional, e → estimated, b → break in time series
lastUpdate	When Eurostat last refreshed the dataset
scrapedAt	Run timestamp

A RUN_SUMMARY record in the key-value store holds per-dataset counts, the cube shape, the filters used, and any dataset that failed.

Common use cases

• EU market analysis and country comparison. Pull one indicator across every member state on one time base.

{
"datasets": ["nama_10_gdp"],
"geo": ["DE", "FR", "IT", "ES", "PL", "NL"],
"timeFrom": "2015",
"filters": { "unit": "CP_MEUR", "na_item": "B1GQ" }
}

• Policy and economic research. Unemployment and inflation on a common panel, ready for regression.

{
"datasets": ["une_rt_a", "prc_hicp_aind"],
"geo": ["EU27_2020", "DE", "FR", "IT"],
"timeFrom": "2010"
}

• Feed a European business dashboard. A small, cheap nightly refresh of headline population figures.

{
"datasets": ["tps00001"],
"geo": ["EU27_2020", "DE", "FR", "IT", "ES"],
"timeFrom": "2020"
}

Quick start

Headline euro-area GDP, current prices, since 2018:

{
 "datasets": ["nama_10_gdp"],
 "geo": ["EU27_2020", "DE", "FR", "IT", "ES"],
 "timeFrom": "2018",
 "filters": { "unit": "CP_MEUR", "na_item": "B1GQ" }
}

Two units side by side, so you can see nominal and real growth together:

{
 "datasets": ["nama_10_gdp"],
 "geo": ["DE", "FR", "IT"],
 "timeFrom": "2020",
 "timeTo": "2024",
 "filters": { "unit": ["CP_MEUR", "CLV10_MEUR"], "na_item": "B1GQ" }
}

Exploring a dataset you do not know yet, grab a capped slice and read the dimensions block to learn its axes:

{
 "datasets": ["une_rt_a"],
 "geo": ["DE"],
 "timeFrom": "2023",
 "maxSeriesPerDataset": 50
}

Input

Field	Type	Default	Notes
datasets	array	-	Required. Dataset codes such as nama_10_gdp. Data Browser URLs are accepted and reduced to the code.
geo	array	[]	Territory codes: countries (DE), aggregates (EU27_2020, EA20), NUTS regions (DE21). Empty = every territory in the dataset.
timeFrom	string	""	Oldest period, in the dataset's own format (2015, 2015-Q1, 2015-01). Empty = full history.
timeTo	string	""	Newest period, same format. Empty = up to the latest release.
filters	object	{}	Any other dimension, as {"unit": "CP_MEUR", "na_item": "B1GQ"}. Values may be a single code or an array.
maxSeriesPerDataset	integer	0	Cap observations saved per dataset. 0 = no limit.
language	string	EN	Label language: EN, FR or DE. Codes are unaffected.

Which combinations make sense. filters is applied to every dataset in the run, so batch datasets that share a dimension vocabulary. nama_10_gdp and nama_10_pc both understand unit and na_item; tps00001 does not define unit at all and will be rejected with INVALID_QUERY_DIMENSION if you send it. That failure is recorded per dataset and the rest of the run continues, but if your datasets do not share dimensions, give each one its own run.

geo, timeFrom/timeTo and filters all apply server-side, so every filter you add makes the download smaller and the run faster. Filtering matters more here than in most APIs: unfiltered, nama_10_gdp returns 1,093,133 observations in an 18 MB response. Adding unit and na_item, the two dimensions that dataset multiplies out most, usually cuts it by two orders of magnitude. If you do not know a dataset's dimensions yet, run it once with maxSeriesPerDataset: 50 and read them off the dimensions field.

Output example

A real record, GDP for Germany in 2023:

{
 "dataset": "nama_10_gdp",
 "datasetLabel": "Gross domestic product (GDP) and main components (output, expenditure and income) - annual data",
 "geo": "DE",
 "geoLabel": "Germany",
 "time": "2023",
 "value": 4219310,
 "unit": "Current prices, million euro",
 "dimensions": {
 "freq": { "code": "A", "label": "Annual", "dimensionLabel": "Time frequency" },
 "unit": { "code": "CP_MEUR", "label": "Current prices, million euro", "dimensionLabel": "Unit of measure" },
 "na_item": { "code": "B1GQ", "label": "Gross domestic product at market prices","dimensionLabel": "National accounts indicator (ESA 2010)" },
 "geo": { "code": "DE", "label": "Germany", "dimensionLabel": "Geopolitical entity (reporting)" },
 "time": { "code": "2023", "label": "2023", "dimensionLabel": "Time" }
 },
 "flags": [{ "code": "p", "label": "provisional" }],
 "lastUpdate": "2026-08-04T23:00:00+0200",
 "scrapedAt": "2026-08-06T11:39:16.686Z"
}

flags is an array because Eurostat concatenates status marks, a bp cell is a break in the series and provisional, and both are decoded separately.

RUN_SUMMARY looks like this:

{
 "datasetsRequested": 2,
 "datasetsFailed": 0,
 "failures": [],
 "seriesSaved": 240,
 "perDataset": [
 { "dataset": "nama_10_gdp", "cells": 240, "observations": 236, "saved": 236, "capped": false }
 ],
 "filters": {
 "geo": ["DE", "FR"],
 "timeFrom": "2018",
 "timeTo": "",
 "dimensions": { "unit": "CP_MEUR", "na_item": "B1GQ" },
 "maxSeriesPerDataset": 0,
 "language": "EN"
 },
 "finishedAt": "2026-08-06T11:39:16.686Z"
}

cells is the size of the cube, observations the number that actually carry data. The gap is normal, Eurostat cubes are sparse, and a country that did not report a given year simply has no cell.

Finding a dataset code

Open the Eurostat Data Browser and pick any table. The code is the last path segment of the URL:

URL you see	Code
…/databrowser/view/NAMA_10_GDP	nama_10_gdp
…/databrowser/view/tps00001/default/table	tps00001
…/databrowser/view/une_rt_a	une_rt_a

You can paste the whole URL, the Actor extracts the code itself. The same page lists every dimension of the dataset and the codes each accepts; those are what go in filters.

Useful starting points: nama_10_gdp (GDP and main components), tps00001 (population on 1 January), une_rt_a (annual unemployment rate), prc_hicp_aind (harmonised inflation), nrg_bal_c (energy balances), ext_lt_intratrd (intra-EU trade).

API example

curl -X POST "https://api.apify.com/v2/acts/arman-bd~eurostat-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
 -H "Content-Type: application/json" \
 -d '{
 "datasets": ["nama_10_gdp"],
 "geo": ["EU27_2020", "DE", "FR"],
 "timeFrom": "2020",
 "filters": { "unit": "CP_MEUR", "na_item": "B1GQ" }
 }'

JavaScript example

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_TOKEN' });
const run = await client.actor('arman-bd/eurostat-scraper').call({
 datasets: ['nama_10_gdp'],
 geo: ['DE', 'FR', 'IT'],
 timeFrom: '2020',
 filters: { unit: 'CP_MEUR', na_item: 'B1GQ' },
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const row of items) console.log(`${row.geoLabel} ${row.time}: ${row.value} (${row.unit})`);

Limits and behaviour

• JSON-stat is dimension-indexed, not tabular. The whole point of this Actor is the unflattening: value is a flat map keyed by a linear index into a cube whose axes are listed in id with sizes in size, last axis varying fastest. The decoder in src/jsonstat.js reverses that arithmetic and asserts the cube's shape before it starts, because a wrong stride order does not throw, it silently attaches every number to the wrong country and year.
• Sparse cubes are normal. A cube of 54 territories may carry 46 values. Missing cells are skipped, never emitted as zero or null.
• Unfiltered queries are big. nama_10_gdp with no filters is 18 MB and over a million observations. The Actor logs a warning when you run with no filters at all, and maxSeriesPerDataset exists as a hard stop.
• A bad dataset code does not kill the run. Eurostat replies HTTP 404 with an explanatory error.label, which is recorded verbatim in RUN_SUMMARY.failures; the Actor only errors out if every dataset fails.
• Invalid dimension names are fatal per dataset, not per run. Passing a dimension the dataset does not define returns HTTP 400 INVALID_QUERY_DIMENSION, recorded the same way.
• Transient errors are retried. 429 and 5xx get three attempts with linear backoff. 404 and 400 are not retried, because they will never succeed.

FAQ

Do I need a proxy? No. Proxy configuration is not required to run this Actor.

Do I need a Eurostat account or API key? No. You supply no credentials.

What happens if one dataset is unavailable? It is logged, added to RUN_SUMMARY.failures, and the run continues with the rest.

Why do I get fewer rows than cells in the summary? Eurostat cubes are sparse. cells counts every possible combination of dimensions; observations counts the ones with a published figure.

How do I know which filters a dataset accepts? Run it once with maxSeriesPerDataset: 50 and read the dimensions object on any returned row, it names every axis and shows an example code for each. The Data Browser page lists the full code lists.

What do the flags mean? They are Eurostat's observation status marks. The most common are p (provisional), e (estimated), b (break in time series), d (definition differs), u (low reliability). The Actor decodes each one to its label as published with the dataset.

Can I get quarterly or monthly data? Yes, pick a dataset that publishes at that frequency and use its period format in timeFrom/timeTo (2023-Q1, 2023-01). The freq dimension in the output tells you what you got.

Can I schedule it? Yes. Eurostat publishes on a release calendar, and lastUpdate on every row lets you skip runs where nothing changed.

Can I integrate it with something else? Yes, Apify API, client libraries, webhooks, scheduled runs, dataset exports (JSON/CSV/Excel) or MCP. Output is structured JSON.