> For the complete documentation index, see [llms.txt](https://docs.talentmap.com/knowledge-base/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.talentmap.com/knowledge-base/step-5-analyze-interpret-and-report/compass-reports/comments/comment-settings/sentiment-analysis.md).
# Sentiment Analysis
### How Sentiment Analysis Works
Compass uses natural language processing (NLP) to automatically categorize open-ended comments as positive, neutral, or negative. This helps you scan themes across large volumes of feedback without reading every comment individually.
#### The Categorization Process
For each comment, the system:
1. Analyzes the language and context of the response.
2. Assigns a sentiment category (positive, neutral, or negative).
3. Generates a confidence score for the categorization.
4. Applies the category only if the confidence score meets a minimum threshold.
Comments below the threshold are left uncategorized rather than risk a misleading label. This is why some comments in your reports may appear without a sentiment tag.
***
#### How to Interpret Sentiment in Survey Data
{% hint style="info" %}
**Sentiment categories work best on clearly positive or clearly negative language.** The underlying model was trained primarily on consumer feedback like product reviews, where comments tend to be unambiguous ("this product is terrible", "I love this service"). Survey comments are often more nuanced.
{% endhint %}
When you ask employees what could be improved, most responses are constructive suggestions rather than strong expressions of positive or negative emotion. A comment like "More cross-team meetings would help us collaborate better" is valuable feedback, but it does not fit neatly into a positive or negative category.
What this means for your reports:
* **A negative skew is common.** Improvement-focused questions naturally surface critical-sounding language, even when the underlying intent is constructive.
* **Neutral does not mean unimportant.** Some of your most actionable comments will be categorized as neutral.
* **Use sentiment as a starting point, not a conclusion.** Sentiment helps you scan and segment comments quickly. The full insight comes from reading the comments themselves, especially within themes you are exploring in action planning.
***
#### What Sentiment Analysis Does Not Do
The current sentiment engine is a traditional natural language processing system, not a generative AI model. It does not:
* Summarize themes across comments
* Identify or extract specific topics
* Interpret nuanced context the way a large language model would
For thematic analysis, use the comments report alongside sentiment categories to identify patterns and group related feedback.
***
### How to Add Sentiment Analysis to a Comment Report
1. Go to **Reports** > **Comments**.
2. Click the **Gear** icon near the top right corner of the page.
3. Check the box for show dimension score, then select a sort option.
4. Click **OK**
{% hint style="info" %}
If running sentiment (for the first time only), the page may take several minutes to analyze all comments. A message will display in the lower left of the page while loading. We recommend refreshing the page after roughly 5 minutes.
{% endhint %}
The page will display an overall sentiment score for all comments. Click through dimensions and items to view the overall score and sentiment scores per comment.
### **Export Comments with Sentiment**
1. Click the **Gear** icon in the right corner then check "Show sentiment score" in preferences.
2. Click Export
The overall sentiment score will show for each item as well as the individual sentiment scores per comment.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.talentmap.com/knowledge-base/step-5-analyze-interpret-and-report/compass-reports/comments/comment-settings/sentiment-analysis.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.