Sentiment analysis

The Label Your Data team can annotate textual data in 42 languages for different industries, including customer support (live chat solutions, bots, etc). We’ve annotated 7000 customer support chats for one of our clients. For this sentiment analysis project, we had to define whether the customer was happy/neutral/sad/angry (i.e., the emotional tone of each message). Our team has also labeled the intent of the customer.

This type of “subjective” data annotation requires using the cross-reference QA method to ensure that no label is biased and that our annotations are of the highest quality and accuracy.