Eurostat

What You're Getting

Eurostat is the statistical office of the EU, providing harmonized data for 27 EU member states plus EEA countries — GDP, employment, trade, agriculture, energy, population, and more. Data is organized in datasets identified by a code (e.g. namq_10_gdp for quarterly national accounts). Each dataset has dimensions (geo, time, unit, na_item, etc.) that you slice with filter parameters.

Ingestion Strategy

Use the SDMX 2.1 REST API or the JSON/Unicode Interchange Format (JSON-UI) API. JSON-UI is simpler to parse. Specify the dataset code and filter dimensions to avoid pulling entire datasets. For bulk extraction, the TSV bulk download is efficient — gzipped TSV files at around 1–100MB each.

import requests

# Quarterly GDP for Germany and France
url = "https://ec.europa.eu/eurostat/api/dissemination/statistics/1.0/data/namq_10_gdp"
params = {
    "geo": "DE,FR",
    "unit": "CP_MEUR",
    "na_item": "B1GQ",
    "s_adj": "NSA",
    "freq": "Q",
    "format": "JSON",
    "lang": "EN"
}
r = requests.get(url, params=params)
data = r.json()

Rate Limits & API Details

• No authentication required; no published rate limit — be polite (1 req/sec for bulk)
• Bulk TSV download: ec.europa.eu/eurostat/api/dissemination/sdmx/2.1/data/{dataset}?format=TSV&compressed=true
• Dataset catalog: ec.europa.eu/eurostat/api/dissemination/catalogue/toc/json

Schema Stability

Dataset codes are stable but Eurostat reorganizes the catalog periodically. Dimension values (geo codes, unit codes) follow SDMX codelists — changes are rare but do happen with EU expansion. Time series for new EU members start at their accession date.

Data Quality Gotchas

• Flags: Eurostat embeds data quality flags (: = not available, b = break in series, e = estimated) in TSV value cells — parse these into a separate column
• NUTS geography levels: data available at NUTS 0 (country), NUTS 1, NUTS 2, NUTS 3 — be explicit about which level you need
• Euro vs national currency: time series for countries before euro adoption show breaks — use PPP-adjusted series for cross-country comparison
• Revisions: quarterly data is revised for 2+ years; store download timestamp

Tool Compatibility

• Python: eurostat PyPI package handles the JSON-UI API and TSV parsing
• R: eurostat CRAN package is mature and widely used
• Airbyte: no official connector
• TSV bulk download works well for Snowflake/BigQuery COPY jobs

Benchmarking Approach

For namq_10_gdp, expect quarterly GDP observations for 30+ countries × 30+ years ≈ 3,000+ rows per unit/indicator combo. Verify Germany Q4 GDP against Destatis press release. After each quarterly release (around 45 days after quarter end), confirm new observations appear. Alert if flag rate exceeds 10% for a country (indicates coverage issues).