The sentiment analysis translates all comments and free-text answers to English before they are analyzed.
The reason for this is to bring various user texts to a unified language (English) for simple keyphrase extraction and sentiment analysis.
Example:
In this example, the free-text answers have been submitted in multiple languages.
To extract the most common and relevant keyphrases, these various texts are translated into English when applying the sentiment analysis, so that the text easily can be analyzed and associated with a sentiment based on a scale of negative-neutral-positive or polarized.