I’m starting a new project and I’m currently in the planning stage.
We’re thinking of deploying the survey to an audience that is spread throughout different countries and we’re expecting people to essentially speak 1 of 3 languages (English being one of them).
How should I approach this?
Should I create 3 separate forms or is there a better way?
I’m concerned about the end game, the Reporting…
You can certainly add additional language support into your form to accommodate different speakers. Have a look at these guides:
And for multi-language XLSForms in general:
Thank you! @Xiphware 
I read both support articles. I prefer XLS form, so that’s the way I will go with.
I will check the ODK link later.
One question:
How do we see the data in Table and Reports inside KoboToolbox’s “Data” area? Is it all together? Or its divided by language?
And the part that confuses me is… how will I deal with the multi-language data, if I have to build an external report? (Excel or Google Looker Studio or Power BI)
Btw, I’ve used Power BI with KoboToolbox before, and it all worked fine, but all the data was single language.
The raw form data in the submission isn’t ‘translated’ in any fashion; it’s just the raw data values. For example, in this select:
The user will be presented - in KoboCollect or Enketo - with either “Yes”/“No”, or “Oui”/“Non” in the UI, depending on their chosen localization. But what will be actually stored and submitted will, regardless of language, in both cases be the raw text (ie name) strings yes or no.
So if you wish to map the raw data back to different languages for, say, external report generator (PowerBi, etc), then you will have to accompany the submission data with the form definition, because it is only the form definition that contains the multi-language translations.
Same is true for other questions, eg text entry or even different calendar types (!). The labels and hints displayed for each question might be translated to different languages for the user, but the data they enter - and submit - is still just the raw text (or in the case of datetimes, ISO 8601 timestamp).
