When uploading data to Canvs, selecting the correct columns for Verbatims (i.e. open-ended text) is critical for accurate analysis. Users sometimes mistakenly import closed-ended data as Verbatims. This can lead to unclear summaries, reduced AI accuracy, and reuploading.
To prevent these issues, Canvs now automatically detects and warns users when a selected Verbatim column appears to function as a Filter rather than true open-ended text responses.
How Filter Detection Works
When users select column(s) containing Responses during data Import, Canvs automatically evaluates the data for:
Low uniqueness: only a few unique values appear in the column.
High repetition: the same values repeated frequently.
If the column(s) selected appears to be a Filter, Canvs displays a warning message when users click Next, prompting them to either select a different column or keep the existing selection as Responses. This warning message acts as a guardrail to prevent common mistakes, while still allowing users to override the warning if the selection is intentional.