NOTE: It is assumed you know how to create custom Dashboards and Charts, and have an understanding of chart Types, Presets, and Metrics. If not, read this article before proceeding.
When building your own Dashboard Charts, you can pull data from various data sources to view different types of analytics events.
In this article, we're going to focus on the Keyword data source, which lets you report on AI‑detected sentiment and usage for specific tracked keywords across your community. It’s one of the sentiment-focused Data Sources, alongside Posts, Post Modifications, and Tags.
The Keyword data source tracks data about:
- the users associated to keywords,
- the Category and/or Discussion where the keyword is located,
- the sentiment score associated with your keywords,
- and more.
Use-case examples
- What are people most often talking about, and how do they feel about it?
- How is sentiment about a specific product/feature trending over time?
- Which keywords are most negative right now (early‑warning radar)?
- Where (which Categories / Groups) is a keyword causing the most negativity?
- Which users and roles are driving sentiment around a given keyword?
- Which keywords are trending more positive/negative compared to last period?
- Did our announcement / fix improve sentiment on a keyword?
- Which tracked keywords are mostly neutral but high‑volume (content opportunity)?
Prerequisites & limits
To get any data in the Keyword Data Source:
- Our Sentiment Analysis & Keyword Tracking feature must be enabled. Learn how.
- Keywords must be tracked on the Sentiment Keywords page.
- Only tracked keywords appear in analytics; untracked ones don’t feed the Keyword Data Source.
- You can track up to 100 keywords at a time, so you’ll want to periodically retire ones that are no longer important to your organization.
Filter Chart data
Controlling what data is shown in a Chart is accomplished via the Group By and Filter options.
- The Group By options (available for pie graphs, line graphs, bar graphs, and tables) enable you to view specific keyword data.
- Use Filters to drill into or exclude specific keyword data.
TIP: You can add one or multiple Group By and Filter options to dissect the data how you see fit. Generally speaking, you'll select Group By options to view a specific slice of data, and, if needed, use Filters to dig deeper.
Let's learn about each of the available Group By and Filter options:
NOTE: These options can be used both to group and filter data.
Available Group By / Filter options
You can group and/or filter Keyword data based on the following event parameters:
- Keyword: The actual tracked keyword string (e.g., Product XYZ) as it appears on the Sentiment Keywords page. Grouping by this shows metrics per specific keyword.
- Keyword ID: The internal numeric ID for the tracked keyword. Functionally similar to Keyword, but useful if you prefer ID‑based joins or need to tell the difference between similarly named keywords (or renamed ones).
- Occurrences: The count of times the keyword occurs in content. For example, when grouped by Keyword, Occurrences is how many times that keyword appeared (in posts/articles) in the selected time range.
- Place Record ID: The ID of the “place” where the content lives – typically a Category or Group.
- Place Record Type: The type of the place referenced by Place Record ID (for example, “category” vs. “group”). Lets you distinguish whether a keyword occurrence was in a category discussion area or a group space.
- Record ID: The unique ID of the content that contained the keyword: discussion, question, idea, comment, or article.
- Record Type: The type of content the keyword was found in, such as discussion, question, idea, comment, article, etc.
- User ID: The ID of the user who authored the content where the keyword is located.
- Sentiment Score: The numeric sentiment value associated with the keyword in its associated context. Lower values indicate more negative sentiment; higher values more positive sentiment.
- User Name: The display name of the user whose content contained the keyword. Lets you group by or filter to specific authors.
- User Rank: The rank name of the user (e.g., “New Member”, “Super User”). Useful for seeing how sentiment varies by engagement level.
- Rank ID: The internal numeric ID for the user’s rank.
- User Role ID: The ID of the user’s role at the time of posting (e.g., specific admin/mod/member role record).
- Role Type: The role type for the user (e.g., admin, cm, mod, member).