Hello, I have some questions about the translation and localization capabilities in GoodData Cloud, particularly concerning the type of data we collect. We handle survey data, so most of the content in our visualizations consists of text like "yes" "no" or other categorical values. In the past, we could only translate the Chrome interface values of GoodData dashboards. To get the data displayed in the customer’s language, we chose to prepare our data before sending it to GoodData—this involved translating values like "yes" and "no" directly into "oui," "non," etc., within our BigQuery dataset. While this approach works, it has become difficult to maintain, especially as we bring on new customers with additional languages. Furthermore, we’ve had to duplicate our visualizations, filters, and dashboards for each language, as LDM requires a different workspace. Recently, I noticed that it’s now possible to export an XLIFF file for an organization in GoodData Cloud. I was hoping that this file would allow me to store translations for the data itself. However, upon exporting it, it only contains filter and visualization names, not the actual data content. Am I mistaken in thinking that the XLIFF file can handle data translations? If so, could you advise on the best approach to translate the data without creating new datasets or duplicating workspaces? Since we embed the dashboards directly in our app, is there perhaps a hook or integration we could use to handle the data translations dynamically? Thank you for your help.

Hi Simon, as you noticed, the XLIFF file in GoodData Cloud helps manage translations for things like dashboard names, filter labels, and visualization titles, but it doesn’t handle actual data translations (e.g., translating data values like “yes” and “no”). When you extend your analytics across various regions and languages, it’s beneficial to adapt the embedded analytics to fit local contexts. This includes translating the content within visualizations, assigning local names to data entities, and adjusting date or currency formats as needed. More details can be found here: https://www.gooddata.com/docs/cloud/customize-appearance/change-display-language/#ChangeDisplayLanguage-MetadataLocalization

@Michael Ullock thanks for having a look at this. I see, XLIFF won't help me translate the actual data then. Do you have any ideas about how to achieve this, i.e. translating the "yes" and "no" values from my database rows in different languages, while keeping a single GC workspace/ldm? Fyi, my main table looks like this, it is a pivot of the questions of a survey:

survey_pivot
- survey_id
- user_id
- questionnaire_id
- date
- question_1
- question_2
- question_...

Hi Simon, You are right, the XLIFF file only translates the UI and Metadata but not the actual data. Unfortunately, there is no easy way how to achieve this without doing some changes in your datasource. You would need to either create some labels for specific languages: https://www.gooddata.com/docs/cloud/model-data/concepts/label/ Or in case that the users shouldn’t see other languages than their own, you can again prepare the values on the datasource level and restrict it via User Data Filters: https://www.gooddata.com/docs/cloud/workspaces/user-data-filters/ However, in case of lots of data and lots of languages used it might cause some performance issues as you would basically duplicate manny entries in your source data. I’m sorry if this was not the answer you were hoping for.

Thanks for your input and time Julius, this definitely helpful. I haven't thought about sending a table with the lines duplicated in different languages, then adding a user filter. In our situation I don't think the performance would be impacted that badly and it would definitely reduce the burden of having fully synchronised workspaces in two or three languages. If we mix that with an xliff with translations of all the chrome, I guess it it could work. Good idea! I will evaluate the tradeoffs of moving our instrumentation towards that, however can you maybe share if GoodData will allow to upload translations for the data at some point? I understand categorical/string values in GoodData is maybe not the priority, however I think it is a missing piece now that we can translate everything else with the XLIFF file. I will also have a look if it's possible to handle that with a dashboard plugin or something. Maybe with the right hook exposed, I could load all translated values beforehand and replace the label with their translations at runtime. Not sure about the side effects though.