evidoria

← Back to browse

Good practice Imported

data.europa.eu's eTranslation — Automated Neural Machine Translation of Open-Data Metadata Across 24 EU Languages

Luxembourg · Luxembourg City · See the Luxembourg profile · See the Luxembourg City profile

Evidence: Descriptive / self-reported Top 6% 87/100 · Ask Evidence Copilot about this practice

The EU's open-data portal auto-translates harvested dataset metadata into all 24 official EU languages via the Commission's eTranslation engine — fully automated, covering 1.6M+ datasets from 178 catalogues — though the portal warns translations may be imperfect.

1.6 million+ datasets
Datasets indexed
178 catalogues
Harvested data catalogues
24 languages
Official EU languages covered
data.europa.eu's eTranslation — Automated Neural Machine Translation of Open-Data Metadata Across 24 EU Languages

Details

Maturity
Established
Promoter
Publications Office of the European Union
Period
2015–present
Region (NUTS)
LU00
Keywords
open data, machine translation, metadata, multilingual government data

Context

The Publications Office of the European Union runs data.europa.eu, the EU's central open-data portal, which harvests metadata from 178 national, regional and EU-institutional data catalogues.

Activities

The portal automatically translates harvested dataset titles and descriptions into all 24 official EU languages using eTranslation, the European Commission's standing neural machine-translation service, which the Commission describes as fully automated with no human intervention.

Results

The portal currently indexes over 1.6 million datasets across its 178 harvested catalogues, all with metadata machine-translated into 24 languages, and the current portal version has run continuously since an April 2021 relaunch that built on a beta operating since November 2015.

Conclusions

Because translation is machine-generated without human review, the portal itself displays a user-facing warning that translations may contain errors, and independent academic research comparing EU and Gulf open-data portals has specifically flagged machine-translated metadata quality as a limitation worth watching rather than a solved problem.

Implementation
Indicative cost
Low (< €50k)
Time to results
Long (> 3 years)
Staffing & skills
The Publications Office of the European Union operates the portal centrally, reusing the European Commission's existing eTranslation neural machine-translation engine rather than running a dedicated translation team.

Conditions for success

  • Reusing a single, already-built, standing translation engine (eTranslation) that is shared across many other EU digital services, rather than building bespoke translation capacity.
  • Continuous institutional ownership by the same body (the Publications Office) since the 2015 beta.

Common failure modes

  • Fully automated translation with no human review means errors can occur, which the portal itself flags to users.
  • Independent academic research has flagged machine-translated metadata quality as an unresolved limitation.

Commonly funded by

Own resources / municipal budget Digital Europe Programme

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful