Analytics & Insights
The Analytics plugin is your built-in business intelligence toolkit. Create interactive dashboards, add visualizations from a rich widget library, connect multiple data sources, and build custom widgets with AI assistance — all without leaving the platform.
Getting Started
Open Analytics from the sidebar. The plugin requires the analytics_access capability on your subscription. You also need an active infrastructure stack provisioned for your tenant.
| Capability | What It Unlocks |
|---|---|
| analytics_access | View dashboards, create and edit dashboards and widgets, access data sources, use the AI custom widget builder |
| analytics_manage | Manage data source definitions, configure metrics |
Interface Overview
The Analytics page is organized into two main tabs:
| Tab | Description |
|---|---|
| Dashboards | Browse, create, view, and edit your analytics dashboards |
| Data Sources | Manage connections to financial, agents, knowledge, and files data |
A stats summary at the top of the page displays key counts — total dashboards, widgets, data sources, and metrics — giving you a quick sense of your analytics footprint.
Dashboards
Dashboards are the primary way to organize and display your analytics. Each dashboard is a responsive grid of widgets that you can arrange, resize, and configure to tell the story your data needs.
Browsing Dashboards
The Dashboards tab lists all dashboards available in your current stack. On smaller screens, dashboards appear as cards; on larger displays, they appear in a sortable data grid with columns for name, description, visibility, widget count, and last update time.
Each dashboard shows whether it is Public (visible to all users in the tenant) or Private (visible only to you). Only corporate administrators and superusers can set a dashboard as public.
Creating a Dashboard
- Click Create Dashboard in the top toolbar or use the floating action button at the bottom-right corner
- Enter a name, optional slug (auto-generated from the name if left blank), and description
- Optionally toggle Public visibility if you have the appropriate role
- Set Auto-Refresh Interval to have the dashboard poll for fresh data on a schedule (in seconds)
- Set Cache Duration to control how long widget data is cached before re-fetching
- Click Create to save and return to the list, or Create & Edit to jump directly into the Dashboard Builder
Opening a Dashboard
Click any dashboard in the list to open it in the Dashboard Viewer — a read-only display of the full grid with all widgets rendered. From the viewer, you can:
- See when data was last refreshed
- Manually trigger a refresh
- Click Edit to switch into the Dashboard Builder
Dashboard Builder
The builder is a full-screen editing environment where you design your dashboard layout and configure each widget.
Layout
- Left sidebar — the Widget Library drawer, organized by category
- Top toolbar — dashboard name, Preview/Edit toggle, Settings, Revert, Save, and Close
- Main canvas — a responsive grid where you drag, resize, and arrange widgets
Adding Widgets
Open the Widget Library from the left sidebar. Widgets are grouped by category:
| Category | Widget Types |
|---|---|
| Metrics | Metric card, Gauge, Linear gauge, Speedometer, Trend indicator |
| Charts | Line chart, Bar chart, Pie chart, Donut chart |
| Advanced Charts | Advanced line, Scatter plot, Radar chart, Bubble chart, Mixed chart, Polar area, Advanced bar, Advanced pie |
| Tables | Data table |
| Specialized | Heatmap, Correlation matrix |
| Custom | Custom widget (with AI assistant) |
Click a widget type to add it to the grid. Drag widgets to reposition them and use the resize handles to adjust their dimensions.
Configuring a Widget
Click the menu on any widget and select Edit to open the Widget Configuration dialog. Configuration is organized into four sections:
- General — widget name, grid dimensions, active/inactive toggle
- Data — select a data source, bind data fields to the widget, configure query parameters. Financial data sources offer a specialized binding panel with company selection and slot formatting
- Style — theme selection, background color, title and legend visibility, grid lines, company branding for financial widgets
- Advanced — auto-refresh interval (0–3600 seconds), caching toggle, custom CSS classes, description
Tip: Use the company branding toggle on financial widgets to display the company logo and ticker symbol alongside your data.
Widget Actions
Each widget offers a context menu with actions:
- Edit — open the configuration dialog
- Duplicate — create a copy of the widget on the same dashboard
- Delete — remove the widget
- Refresh — manually re-fetch the widget's data
Saving and Reverting
- Click Save to persist all layout and widget changes
- Click Revert to discard unsaved changes and restore the last saved state
- If you close the builder with unsaved changes, a confirmation dialog gives you the option to Discard, Cancel, or Save & Exit
Preview Mode
Toggle Preview in the toolbar to see the dashboard exactly as end users will. The widget library sidebar hides and all editing controls are disabled so you can evaluate the final layout.
Custom Widgets with AI Assistant
Custom widgets are a standout capability of the Analytics plugin. They let you write your own rendering logic using JavaScript, with the option to have the AI assistant generate or refine the code for you.
How Custom Widgets Work
A custom widget executes your JavaScript code in a sandboxed environment. You write rendering logic using React.createElement (no JSX), and the widget has access to the data bound through the standard data source configuration. This lets you build highly specialized visualizations that go beyond the built-in chart and metric types.
Using the AI Assistant
The AI assistant can generate custom widget code based on your plain-language instructions:
- Add a Custom widget to your dashboard from the Widget Library
- Open the widget's configuration and navigate to the code editor
- Enter your instructions describing what you want the widget to display — for example, "Create a dashboard card showing total revenue with a month-over-month comparison and a sparkline trend"
- Select a model type:
- Smart — more capable model for complex visualizations with detailed logic
- Fast — quicker generation for simpler widgets
- Click Generate to have the AI produce the widget code
The AI receives context about your data schema, sample data, and the available rendering API so it can produce code that works with your actual data shape.
Refining Generated Code
After generation, you can iterate:
- Edit the generated code directly in the code editor
- Enter additional instructions and click Refine to have the AI modify the existing code based on your feedback
- Use the Debug accordion to inspect errors and execution details
- Toggle the Settings accordion to adjust the AI model preference
Tip: Start with a simple instruction and refine iteratively. The AI retains context from the current code, so each refinement builds on the previous result.
Custom Widget Use Cases
- Specialized financial metric cards with conditional formatting
- Multi-KPI summary panels combining data from different sources
- Interactive data visualizations tailored to your domain
- Branded report components with custom styling
- Dashboard widgets that combine data from agents, knowledge, and file analytics
Data Sources
The Data Sources tab manages the connections that feed data into your widgets. Data sources are organized into four categories, each with its own management panel:
Financial Data Sources
Connect to market data, company financials, and economic indicators. Financial data sources support multiple query types:
- Price data — real-time and historical stock prices
- Fundamentals — balance sheet, income statement, cash flow
- Historical data — time-series financial metrics
- Earnings — earnings reports and estimates
- Dividends — dividend history and yield
- Market cap — company valuation metrics
- Sector/Industry — sector performance and comparisons
- Economic indicators — macroeconomic data points
Each financial data source can be tested directly from the management panel to verify the connection and preview returned data.
Agents Data Sources
Pull analytics from your AI agent executions — run counts, success rates, token consumption, and execution durations. Use these to monitor agent performance and resource usage over time.
Knowledge Data Sources
Access metrics from your knowledge base — document counts, ingestion activity, category distribution, and search patterns. Useful for tracking how your knowledge library grows and is used.
Files Data Sources
Track file storage analytics — upload activity, storage usage, file type distribution, and processing status. Ideal for monitoring document throughput and storage consumption.
Setting Up Data Sources
When you first open the Data Sources tab, the platform automatically provisions a default set of data sources for your stack. You can also:
- Create new data source definitions with custom connection and query configurations
- Test a data source to validate its connection and preview sample data
- Edit connection parameters, query settings, and caching rules
- Deactivate data sources you no longer need without deleting them
Dashboard Settings
Dashboard-level settings control behavior across all widgets:
| Setting | Description |
|---|---|
| Name | Dashboard display name |
| Description | Optional summary of the dashboard's purpose |
| Public | Whether all users in the tenant can view the dashboard (admin/superuser only) |
| Auto-Refresh Interval | How often (in seconds) all widgets poll for new data; set to 0 to disable |
| Cache Duration | How long widget data responses are cached before re-fetching |
Access dashboard settings from the Settings button in the builder toolbar or from the create/edit dialog.
Best Practices
- Start with a plan — sketch out which metrics and visualizations you need before building. Group related widgets into a single dashboard for a coherent narrative
- Use built-in widget types first — the standard chart, metric, and table widgets cover most visualization needs and require no code
- Leverage the AI assistant for custom widgets — describe what you want in plain language rather than writing code from scratch. Iterate with the Refine feature to fine-tune the result
- Connect the right data sources — bind financial widgets to financial data sources for real-time market data, and use agents/knowledge/files sources for operational analytics
- Set appropriate refresh intervals — auto-refresh keeps dashboards current, but very short intervals increase load. Use 60–300 seconds for most use cases
- Use public dashboards for team visibility — share key dashboards with the entire tenant so stakeholders have a common view. Keep experimental or work-in-progress dashboards private
- Test data sources before binding — use the test feature in the Data Sources tab to verify a source returns expected data before wiring it into widgets
Troubleshooting
Analytics page shows a capability warning
Problem: You see a message that the analytics capability is required.
Solution: Contact your administrator to enable analytics_access on your subscription.
Cannot create or edit dashboards
Problem: The Create Dashboard button or Edit action is not available.
Solution: Verify that your role includes the necessary permissions and that analytics_access is active on your subscription. Developer tenants restrict dashboard creation to the user's own resources.
Widgets show no data or mock data
Problem: A widget displays placeholder or random-looking values. Solution: Ensure the widget is bound to an active data source. Open the widget configuration, check the Data section, and verify that a data source is selected and the field mappings are correct. Use the test feature in Data Sources to confirm the source returns data.
Dashboard does not auto-refresh
Problem: Widget data appears stale even though auto-refresh is configured. Solution: Auto-refresh is disabled while the dashboard is in edit mode or preview mode. Open the dashboard in the viewer (not the builder) to activate auto-refresh. Verify the auto-refresh interval is set to a value greater than 0 in the dashboard settings.
Custom widget code errors
Problem: A custom widget shows an error instead of the expected visualization.
Solution: Open the widget configuration and expand the Debug accordion to view the error message. Common issues include undefined variables, incorrect React.createElement syntax, or referencing data fields that do not exist in the bound data source. Use the AI Refine feature to ask the assistant to fix the issue.
Cannot set a dashboard as public
Problem: The public toggle is disabled in the dashboard creation or settings dialog. Solution: Only corporate administrators and superusers can set dashboards as public. Developer tenants cannot create public dashboards. Contact your administrator if you need to share a dashboard across the organization.
Related Topics
- AI Agents — User Guide
- Knowledge Management — User Guide
- Files Management — User Guide
- Dashboard Navigation Guide
⏱️ Read time: 14 minutes | 📊 Difficulty: intermediate | 🔄 Last updated: 2026-03-30