Insight Panel
The Insight Panel Widget is a powerful tool for automatically analysing your data and surfacing meaningful findings — called insights — directly in your report. It works as a smart assistant that continuously scans your data according to rules you define, and presents relevant observations organised into categories.
Each insight can include rich content (text, charts, tables), carry a severity and priority level, and suggest follow-up actions. When connected to other widgets such as a network visualisation, it can show context-specific insights for whichever item (e.g., a person or node) is currently selected.
There is no autosave feature for Reports. Click Save regularly while working in the Report Workspace.
How Insights Are Generated
The Insight Panel works through rules. Each rule defines:
- What data to look at — which data source(s) to pull from.
- When to surface an insight — a condition that determines which data rows produce an insight.
- How severe it is — a severity score between 0 and 1 reflecting how significant each finding is.
- What to show — the title, visual content (text, charts, tables), and suggested actions.
When you save a rule, the system processes every row (or group of rows) in your data, evaluates the condition for each, and creates an insight for every row that passes. All insights from all rules are collected and displayed together in the widget.
Managing Rules
Before the Insight Panel can show any insights, you must create at least one rule.
- Add a Rule — In the Insight Panel tab of the widget settings, click Add Rule. This opens a four-step rule configuration wizard.
- Edit a Rule — Click the settings icon next to an existing rule to reopen its configuration dialog.
- Delete a Rule — Click the delete icon next to a rule. A confirmation dialog appears before the rule and all its generated insights are removed.
Deleting a rule also removes all insights generated by that rule, including any bookmarks, priority changes, or resolution statuses that users had set on those insights.
Configure a Rule
The rule configuration dialog guides you through four steps. You must complete each step's required fields before proceeding. The stepper on the left shows your progress.
Step 1: Data
This step defines where your data comes from, how rows are identified or grouped, what conditions must be met for an insight to be generated, and which field links insights to selected items in connected widgets.
The screen is split into two halves:
- Left side — Data source definitions (primary and secondary).
- Right side — Condition expression, severity expression, identifier attribute, and a preview of the resulting data.
Primary Data
The Primary Data source is required and acts as the foundation of the insight rule.
- Data Source — Select the dataset to use as the primary source.
- Filters (optional) — Apply filters to limit or refine the records included:
- Click Add Filter.
- Field — Select a field from Node Meta, Node Attributes, or Node Analytics.
- Operator — Select the filter operator.
- Value — Select the filter value.
- To delete a filter, click the delete icon on the filter row.
- ID Attribute — Select an attribute that uniquely identifies each record. Use this when each row already represents a distinct item and you want to generate one insight per row.
- Grouping Attribute — Select an attribute used to group records based on shared values. Use this when you want to combine multiple rows into summary rows before generating insights.
- Aggregation Method — Only relevant when using a Grouping Attribute. Controls how numeric values are combined when multiple rows are merged into one group.
ID Attribute vs. Grouping Attribute — These two options are mutually exclusive. When an ID Attribute is selected, each node is treated as an individual record. When a Grouping Attribute is selected, nodes are grouped by the selected attribute value. Setting a Grouping Attribute automatically clears the ID Attribute.
When a Grouping Attribute is set on the primary data, the system automatically sets the primary join attribute for all secondary data sources to match the grouping field. If you change or clear the grouping attribute, the join attribute on all secondary data sources is reset and you will need to re-select it.
Secondary Data
Secondary Data sources are optional datasets that can be joined with the Primary Data to provide additional context and attributes.
To add a Secondary Data Source:
- Click Add in the Secondary Data section.
- Configure the following fields:
- Data Source — Select the data source to bring in.
- Join Type — How to combine rows from this source with the primary data (see Join Types below).
- Join Attributes — Two fields that define the matching key:
- Primary Data Attribute — The field from the primary data to match on. If the primary data uses a Grouping Attribute, this is automatically set to the grouping field and cannot be changed.
- Current Data Attribute — The field from this secondary data source to match on.
- ID Attribute / Grouping Attribute / Aggregation Method — Same concept as primary data.
- Alias — A short name for this secondary data source (e.g.,
ds_1,survey_data). Used to reference this source's fields in expressions and titles. Aliases must be unique, cannot be "primary" (reserved), and spaces are replaced with underscores.
To remove a Secondary Data Source: Click Remove at the bottom of the secondary data entry.
A secondary data source is considered invalid (shown with a warning icon) if any of the following are missing: data source selection, ID or grouping attribute, both join attributes, or alias.
Join Types
The Join Type determines which rows appear in the final result when combining primary and secondary data.
| Join Type | Description |
|---|---|
| Inner Join | Only keeps rows where a match exists in both datasets. Unmatched rows from either side are excluded. |
| Left Join | Keeps all primary data rows. Matched secondary rows are combined. Unmatched primary rows are kept with empty secondary fields. |
| Right Join | Keeps all secondary data rows. Matched primary rows are combined. Unmatched secondary rows are kept with empty primary fields. |
| Full Join | Keeps all rows from both datasets. Matched rows are combined. Unmatched rows from either side are kept with empty values for the missing side. |
If one primary row matches multiple secondary rows (or vice versa), each combination creates a separate result row. For example, if one person (primary) has three network relationships (secondary), the join produces three rows for that person.
Condition Expression
The Condition Expression determines which rows should produce an insight. It should evaluate to true or false — only rows where the condition evaluates to true generate an insight. The expression editor provides autocomplete suggestions showing available fields from all configured data sources.
Fields are referenced using a path format:
primary.data.fieldName— a field from the primary data source.primary.analytics.relationship.metric— an analytics metric from the primary data source.ds_1.data.fieldName— a field from a secondary data source with aliasds_1.
Examples:
primary.data.score > 80— Only generate insights for rows where the score field is greater than 80.primary.analytics.friends.in_degree > 5— Only for rows where the in-degree on the friends network is greater than 5.primary.data.department == "Sales" && ds_1.data.performance < 50— Combining conditions from two data sources.
If you leave the Condition Expression empty, every row in the data will generate an insight.
If the expression has a syntax error or references a non-existent field, an error message appears below the editor and you cannot proceed to the next step.
Severity Expression
The Severity Expression assigns a numeric severity score (between 0 and 1) to each insight, indicating how significant or urgent a finding is. The expression works the same way as the Condition Expression but should produce a number between 0 and 1. Values outside this range default to 0.
Severity Thresholds divide the 0–1 range into four levels:
| Threshold | Level | Default Range |
|---|---|---|
| Below Low Threshold | No Severity | below 0.40 |
| Low to Mid Threshold | Low Severity | 0.40 – 0.60 |
| Mid to High Threshold | Medium Severity | 0.60 – 0.80 |
| Above High Threshold | High Severity | above 0.80 |
You can adjust the three threshold values. They must be in ascending order (Low < Mid < High).
Severity is displayed as a coloured label on each insight: No Severity (neutral), Low Severity (subtle), Medium Severity (yellow/warning), High Severity (red/danger).
Identifier Attribute
The Identifier Attribute enables the Insight Panel to react to selections in other connected widgets. It consists of two parts:
- Source Data — Choose which data source the identifier comes from: Primary or a secondary data source (listed by alias).
- Source Data Attribute — Choose the specific field within that data source.
When the Insight Panel is connected to another widget (e.g., a network visualisation), selecting an item in that widget sends a selected identifier value to the Insight Panel, which then filters its insights to show only those whose Identifier Attribute value matches the selected item.
The Identifier Attribute is required. Even if you do not plan to connect the Insight Panel to another widget, you must set it. In that case, select any meaningful field — all insights will be displayed together since no selection-based filtering will occur.
Insights whose identifier value does not match any selection are placed in a general group and shown when nothing is selected.
Resulting Extracted Data Preview
At the bottom-right of the Data step, a data grid shows a preview of the final joined data. Use this to verify that your data sources are connected correctly, the join produces the expected number of rows, required fields are present, and values look correct after any grouping or aggregation.
If the configuration is incomplete, the grid shows: "Configure data sources to see results."
Step 2: Properties
This step defines the metadata for each insight generated by this rule — how it appears, how it is labelled, and how it is organised.
Insight Title
The Insight Title is the heading users see for each insight. It is required. The title supports dynamic data fields — type [ to see available fields from your data sources. Selecting a field inserts it as a placeholder (e.g., [primary.data.name]). When insights are generated, each placeholder is replaced with the actual value from that row.
Example: A title of [primary.data.name] has a low engagement score produces insights like "Alice has a low engagement score" and "Bob has a low engagement score".
Rule Name
An optional label for this rule, visible only in the designer. It does not appear in the published report.
Priority
Sets the default priority level for all insights generated by this rule:
- None — No special priority.
- Low — Low importance.
- Medium — Moderate importance.
- High — High importance.
Priority is displayed as a coloured indicator on each insight. Users can also change the priority of individual insights at runtime.
Important Flag
When enabled, all insights from this rule are marked as important and automatically placed in the special "Important" category near the top of the categories list.
Categories
Assign one or more categories to organise insights generated by this rule. You can select from existing categories or create new ones by typing a name and pressing Enter. Multiple categories can be assigned to a single rule — insights will appear in each assigned category.
Reserved category names (managed by the system):
- Bookmarks — Automatically populated with bookmarked insights.
- Important — Automatically populated with insights flagged as important.
- Resolved — Automatically populated with resolved insights.
Tags
Tags provide an additional way to label and filter insights. Like categories, you can select from existing tags or create new ones. Tags are used for cross-widget filtering — when the Insight Panel is connected to a tag-controlling widget, users can filter insights by selecting specific tags in that connected widget.
Step 3: Content
This step defines what users see when they click on an individual insight. Each insight can contain one or more content items displayed in the order you arrange them.
Content Page Layout
- Content List (left) — Displays all content items added to the insight. Create, edit, delete, rename, and reorder items from here.
- Content Preview (right) — A live preview of how the selected content appears within the Insight Panel.
Content Management
Add content:
- From the top-right of the page, select a content type (Text, Chart, or Table).
- Click + Add — the content configuration page opens.
Edit content:
- From the Content List, locate the item and click Edit.
- Update the required settings.
- Click Confirm to save.
Delete content: From the Content List, locate the item and click Delete.
Reorder content: Click the Up or Down arrow next to the content name in the Content List.
Rename content:
- From the Content List, click Edit for the item.
- At the top of the configuration page, click the Edit icon next to the current content name.
- Enter the new name and click the confirm icon to save.
Content items are displayed to users in the same order shown in the Content List.
Content Types
- Text — A rich text block. Supports data field mentions — insert dynamic values using the
@mention system. When insights are generated, mentions are replaced with actual values from each row's data. For example: "This person has a score of[primary.data.score]which is below the threshold" becomes "This person has a score of 42 which is below the threshold" for a specific row. - Chart — An embedded chart (line, bar, pie, scatter, radar, etc.) with its own data mappings. Charts can use content filters to automatically filter data based on the current row's values, making each insight's chart context-specific.
- Table — An embedded data table with configurable columns. Like charts, tables can use content filters for row-specific data. Additional display mode options:
- Fit Content — the table grows to show all rows (with optional pagination).
- Fit Parent — the table fills the available space.
- Fixed Height — the table has a specific pixel height.
Content Filters
Content filters allow each chart or table content item to dynamically filter its data based on the current insight's row data. Each filter has four fields:
- Current Data Attribute — The field from the content's data source to filter on.
- Operator — The comparison method (equals, not equals, greater than, less than, etc.).
- Source Data — Which rule data source contains the value to compare against (primary or a secondary alias).
- Source Data Attribute — The specific field from that source whose value (from the current row) is used as the filter value.
Sample Data Preview
The Content step includes a sample data row preview at the top-right. Use the left/right arrow buttons to cycle through rows of your extracted data — the preview updates to show how the content will look for each specific row, helping you verify dynamic content before finalising the rule.
Step 4: Actions
The Actions step allows you to define follow-up actions that users can take based on an insight. Actions are optional — if insights are purely informational, you can skip this step and click Confirm to finish.
Add an Action:
- On the Actions page, click Add Action — a new action entry is added.
Remove an Action: Click Remove for the action you want to delete.
Each action requires the following:
- Title — A short name for the action. Supports dynamic data fields using brackets
[...]. The title appears in the action list. - Description — A detailed explanation of the action using a rich text editor. Also supports dynamic data fields. Shown when a user clicks on an action to expand it.
Both the title and description are required for each action. If either is empty, the rule cannot be confirmed.
Using the Insight Panel Widget
Once rules are configured and insights are generated, the widget presents them in a rich, navigable interface.
Categories View
The default view when the widget loads. It shows all categories as a list of cards. Each card displays the category name, insight count, status breakdown (pending vs. resolved), and priority distribution.
Special categories appear in a fixed order:
- Bookmarks (if any insights are bookmarked) — at the top.
- Important (if any insights are flagged as important) — second.
- Your custom categories — sorted alphabetically.
- Resolved (if any insights are resolved) — at the bottom.
A View All button switches to a flat list of all insights.
All Insights View
Shows every insight as a flat list without category grouping. Each insight card shows the title, a coloured priority indicator (green for low, yellow for medium, red for high, grey for none), and a resolved indicator if applicable. A View Categories button returns to the categories view.
Category Details View
Clicking a category shows all insights belonging to it, scoped to that category. A back button returns to the categories list.
Insight Details View
Clicking an insight opens its full detail view, which includes:
- Back button — returns to the previous view.
- Bookmark button — toggle the insight's bookmarked status.
- Resolve button — mark the insight as resolved or un-resolve it.
- Priority selector — change the priority (None, Low, Medium, High) with colour coding.
- Severity label — shows the severity level with colour coding and a tooltip showing the exact percentage.
- Title — the insight's title text.
- Content — all content items (text, charts, tables) rendered with actual data values for this specific insight.
- Suggested Actions — if actions were defined in the rule, they appear at the bottom. Click an action to see its details and assign it.
Sorting Insights
Click the sort icon in the widget header to open the sort menu. Sort by:
- Priority — High to Low, or Low to High.
- Severity — High to Low, or Low to High.
Selecting an already-active sort option clears the sorting and returns to default alphabetical order.
Filtering Insights
Click the filter icon in the widget header to open the filter menu:
- By Insights — Actionable (insights with actions) or Not Actionable (insights without actions).
- By State — Resolved or Not Resolved.
- By Priority — None, Low, Medium, or High. Multiple selections are supported.
Click Apply to activate selections. Click Clear All to remove all filters. When filters are active, the view resets to the categories view.
Searching Insights
The search bar below the header lets you search insights by title. As you type, a dropdown shows matching insights. Click a result to jump directly to that insight's detail view.
Bookmarking and Resolving Insights
- Bookmarking — Bookmarked insights appear in the Bookmarks category. Use this to mark insights you want to revisit.
- Resolving — Resolved insights move out of their regular categories into the Resolved category at the bottom.
These statuses are preserved when rules are re-evaluated, as long as the underlying data still produces the same insight.
Working with Actions
When you open an insight that has suggested actions, they appear as a list labelled Suggested Actions at the bottom of the detail view. Clicking an action opens an action detail panel where you can:
- Read the action's description.
- Assign the action to one or more team members from your project.
- Set a due date.
- Submit the assignment.
Connecting to Other Widgets
Responding to Selections
When the Insight Panel is connected to a widget that sends selection events (such as a network visualisation), clicking an item in that widget causes the Insight Panel to automatically filter its insights to show only those whose Identifier Attribute value matches the selected item. Clicking away from any item (deselecting) shows all insights again.
Tag-Based Filtering
The Insight Panel also supports tag-based filtering. When connected to a tag-controlling widget, it shows only insights whose tags match the selected tags, enabling flexible cross-filtering between different parts of your report.