A user would like to specify a list of categories to force into the list of categories created when using Displayr AI for text analysis category creation. From the user:

"I would like to add a custom list to match to and/or copy from a different project in the future."

When customers use the method, there's no easy way to adjust the coding. If they additionally use this method via an existing categorization, there's no guarantee that it will classify everything. To continue coding, a user would need to create a filter of respondents not coded, filter the text variable(s) by this filter, then perform back-coding. It would be great to streamline this process better so it can be done more easily without multiple steps.

A customer recently mentioned that one of the biggest concerns is the ability of using Displayr AI in surveys with existing code frames based on their back data (i.e. Brand Health Trackers).

They would like to ensure that Displayr AI takes an existing code frame into account and:

This will SIGNIFICANTLY improve a lot of customers' workflow in Tracking studies.

Users often will want to use semi-automatic or manual coding to code a partial dataset, then use the coded portion of the dataset and use Automatic Categorization to extrapolate the rest.

This process requires two separate interfaces.

The suggestion here is to build one interface that will allow the user to:

Code just a portion of their data (“train” it)

Then apply the coding to the rest of the data