The Matrix view, located in the "Explore" section, helps you synthesize relationships and summarize relevant content between two sets of user-defined variables in your data. It offers a visual approach to uncovering insights from complex datasets.
First, define the row and column variables you want to use in the matrix.
For this example, we’ll perform a PESTEL analysis on developments related to electric vehicles in a few selected countries. We’ll set each PESTEL factor as a column and each country of interest as a row.
We begin by naming our variables as "country" for the rows and "PESTEL" for the columns.
Next, add specific row and column headers by clicking the plus symbol. The symbol next to the plus sign allows you to use metadata as a variable. To remove a variable, click the "X" symbol.
You can also summarize all content in a row or column by enabling the “Summarize rows” and “Summarize columns” options.
To switch the orientation of your data, use the “Transpose” button to swap rows and columns.
Once you’ve inserted all the variables, write the instruction using the variables you’ve defined. Your instruction should clearly specify what the AI needs to do with the variables to generate the desired output.
Alternatively, you can write a custom instruction and select specific Knowledge Bases for either row or column values by expanding the variable using the downward arrow.
Please note that you can add custom instructions for rows or columns, but not both at the same time.
In the example below, we defined what we meant by “Political” and selected only one Knowledge Base. As a result, the “Political” column in the matrix will reflect that specific instruction and data source.
For more complex tasks, you can switch to a more advanced LLM engine, which may incur higher costs. Otherwise, use the default LLM.
After defining your variables, prompt, and selecting the desired LLM engine, you’re ready to proceed. Scroll to the bottom of the sidebar and click on the “Save” button to confirm your setup. Once saved, the system will begin generating the matrix. Please allow some time for it to load.
Once the matrix loads, it will appear as shown below. Empty intersections between rows and columns indicate that no relevant data was found in your Knowledge Bases for that specific combination. To explore the data, hover over or click on a bubble.
Once you click a bubble, a text card will open, displaying the output text and specific references from your data.
To download your matrix, click the "Download as Excel" icon located in the top left corner.
To publish your matrix, use the icon to the right of the download icon.
Share settings
Invite members: Type the email address of a teammate in your organization. Use the dropdown to select their role (e.g., Viewer or Editor), then click Add.
Members with access: This section shows who has access. If you have created the Insight Booster project and have not shared with anyone yet, you should see your email address only.
General access: Set the visibility of the matrix for others i.e. ‘Restricted’ vs. ‘Anyone with link’. Restricted option will only share the project with the team members you explicitly add; and ‘Anyone with link’ will grant access to anyone with link.