DataHub Now Speaks Your Language: Introducing Multi-Language Support
The DataHub community has always been global. Nearly half of users are outside the United States, and the majority of DataHub Cloud customers have operations that span multiple countries. Our Ask DataHub agent can answer questions in most languages already, and nearly 20% of our DataHub Cloud customers have had at least one user ask DataHub a question in a language other than English!
Although we increasingly rely on AI agents for analytics, humans are still the most important users of data inat any organization. For our customers in non-English speaking countries, having the key interface to an organization’s data not available in the everyday language of the business creates friction. That’s why we’ve integrated the i18n language localization framework into DataHub, which enables multi-language support in both DataHub Core and DataHub Cloud. You can find more details in the DataHub Docs.
New languages in DataHub Cloud
Multi-language support defaults on and follows the browser locale; German (GA), Spanish, Brazilian Portuguese, French, Italian, Norwegian, Swedish, and Hungarian locales are available (Private Beta).
To choose a new language, go to Settings → Appearance → User interface language and pick your language from the dropdown menu. This translates all buttons, labels, menus, and error messages into your target language instantly, and AI search understands your new language already. Language is customizable at the user level, enabling users at the same organization to see the DataHub interface in the language of their choice.
If you don’t see this option, talk to your account team to make sure you have this feature enabled.
As always, Ask DataHub is still a multilingual guide to your data. You can ask questions in nearly any language, irrespective of the interface language you choose.
If your team needs language that is not on this list, tell us. Reach out to your account team or open a discussion in the DataHub Slack.
How we did it
To create a framework for localization, we built a language registry for DataHub, and gave every user-visible string a reference so it can be rendered in any language assigned by that registry. We then developed a style guide that defines whether those strings should stay in English or be translated. For example, depending on language, some terms, like “tag” or “analytics” should stay in English.
To add a language, a developer can add it to the registry along with a translation of the labeling file. Of course, LLMs are what make the labeling a much faster process. With that framework in place, an LLM takes a first stab at translating the labeling file, and a native speaker reviews and refines the result.
A call to the open source community
Everything described above is also available in open source DataHub, and the framework is designed for community contributions. If you want to add a language we have not shipped yet, you can add it yourself.
If you want to submit a language, the recently merged PR for French language support is a good blueprint to follow. You can also search the codebase for “i18n” to find everything related to internationalization, and join the #design-i18n channel in DataHub Slack to ask questions or share progress.
And, please help us refine our translations! Although we are lucky at DataHub to have talent all over the world to review translations in many languages, this is where we can use your help. As you find places where our translation is not ideal, feel free to submit a PR, or just reach out directly in Slack or to your account team and share the feedback. This helps us refine terminology and context so the translation reads the way a native speaker would expect.
Review recent language support PR examples:
A word of thanks
Finally, a word of thanks. This has been a perennial topic in the DataHub OSS community, and we appreciate the work of the members who found ways to localize DataHub even before LLMs made it easy. For example, community member Luis Hernandez Salazar developed a solution to translate the BoaVista Platform, a major cloud-native data and credit bureau platform, into Brazilian Portuguese. (You can read more about how he did it in this Medium article.) Community members Naiche Despouy of Datanalysis Reunion in France and Tim Schlitzer of it-novum in Germany (who’s now joined the DataHub team) also made contributions that were instrumental in kicking off the project.
DataHub has always been built by a global community. Whether you’re refining a translation, adding a language we haven’t shipped yet, or just telling us what your team needs, we’re glad to have you helping make DataHub speak everyone’s language.
