What is taxonomy scoring?

Taxonomy scoring allows you to assign or calculate a custom sentiment score for specific taxonomy labels. Sentiment scores are calculated (and aggregated) solely with records falling within a particular label, allowing for more distinct and accurate insight into customer sentiment. This is distinct from filtering the sentiment scores from a topic or neural sentiment model with taxonomy model outputs. By evaluating verbatims at the granularity of an individual classification topic, you get significantly more accurate, actionable insights to help improve your overall customer experience.

Availability

Taxonomy scoring is available for all accounts.

How do I add scoring to an existing taxonomy?

Open the taxonomy editor for the desired taxonomy and click the label you wish to apply a sentiment score.
Once in the label, click "+ Set score" in the top-right of the editor. You can assign a custom score or a sentiment model to the label.
You'll see the sentiment label appear when it has been applied.
Save the updated taxonomy and navigate to your workspace.
Proceed to deploy the analyses similar to how you would normally deploy a taxonomy analysis. However, in the “Choose Taxonomy Type” dropdown menu, please select “Taxonomy with Scoring”.
Once the analysis has finished processing, you can access the score output through the widget editor. Choose the data field “taxonomy.score” under the taxonomy model results, along with any other desired categorical or taxonomy-related labels.
Save your widget, and adjust the aggregation of the output as needed.
Once back to your workspace you will be able to see the score and changes dynamically in your workspace.