The Sentiment Analysis tool uses AI to assign a Positive, Neutral, Mixed, or Negative tone to text responses in Text Analysis. Combined with tags and demographic filters, this will help you understand the sentiment your community has towards the project.
Admins can see a Sentiment Summary chart, correct or confirm the sentiment, and refine your analysis using tags or filters. Sentiments will also be available in your downloaded report, and you can use sentiment to gain some quantitative insight from open-ended responses.
Please note that Sentiment Analysis is not available with the Quick Poll or Questions tools, and we can only support English and French languages. Additionally, the AI we use for Sentiment Analysis doesn’t learn and can be fallible. It’s important always to review the sentiment given to ensure accuracy.
To use Sentiment Analysis:
Go to Reporting > Text Analysis, choose your project, tool, and time frame, and select Update.
Use the Sentiment Analysis icon to ensure the feature is on.
The Sentiment Summary will show an overall percentage of what sentiments are present. You can select one to filter your responses to that sentiment.
Each comment will show the sentiment, a confidence score, and a box you can check to Confirm Sentiment.
To change a sentiment, choose a different one using the sentiment dropdown menu. Once you select it, the Confirm Sentiment box is checked; uncheck it to edit the sentiment again.
If you have changed a sentiment, Update the Sentiment Summary to ensure it is accurate.
At any time, you can add tags or use the filters and search to narrow your list of responses. You can also select Download Report to download an Excel file containing each comment and the sentiment and tags applied.
