View usage and feedback metrics from Customer Experience Insights over time with a configurable dashboard. Configurable dashboards offer flexibility in data representation and layout because you can build and maintain custom views with a variety of charts. The configurable dashboard gives you the ability to select a chart type, store the configuration, and retrieve it the next time you visit the page.
Dashboard components
A dashboard contains the following components.
| Component | Description |
|---|---|
| Dashboard | The top-level resource that stores the entire structure and chart definitions. |
| Section | A section can have tiles or nested sections (these appears as tabs). Dashboard sections can have a display name and description. You can apply the date_range filter to all the charts in an entire tab. Sections define their physical dimensions using width and height in grid units. |
| Tile | An element within the dashboard, defined as a nested section or a chart. The default date range is last week. You can further query the date_range as last 1 day, last 2 days, last 1 week, last 2 weeks, last 1 month, last 1 quarter, last 1 year or the customized start and end dates. |
Dashboard field definitions
Each dashboard is made up of the following fields. Only display_name and description appear in the console.
| Field | Description | Constraints/Notes |
|---|---|---|
name |
Dashboard name | Format: projects/{project}/locations/{location}/dashboards/{dashboard} |
display_name |
User-provided name for the dashboard | Maximum of 100 characters |
description |
Dashboard description | Optional |
create_time |
Dashboard creation timestamp | Output-only |
update_time |
Dashboard last update timestamp | Output-only |
filter |
Filter for all charts in the dashboard | Supports optional string filter |
Data and querying
The chart uses natural language (NL), which Customer Experience Insights translates in SQL queries. These queries generate the data and visual specifications to display the chart. Visual specifications are in a VEGA-LITE format. The chart API uses the GenerativeInsights data source, session_id, and revision_id to store the current SQL and visual specifications states of a chart.
Dashboard operations
You can perform the following operations on configurable dashboards.
Create dashboard
Follow these steps to create your dashboard.
Console
Navigate to the CX Insights console, sign in with your Google Account, and select your project.
Click Dashboard. Here, you can see both prebuilt and custom dashboards.
On the Dashboard studio, click + Add dashboard to create your own custom dashboard.
Each dashboard appears in the form of a tile.
When you click a dashboard, it gives an overview of all the charts.
REST API
Submit a request like the following example.
{
"displayName": "My New Dashboard with Container",
"root_container": {
"displayName": "Overview tab",
"widgets": [
{
"container": {
"displayName": "Inner Chart Group 1",
"widgets": [
{
"chart": {
"displayName": "Most Reported Issues"
"dataSource": {"generativeInsights": "What are top 5 issues?"}
}
},
{
"chart": {
"displayName": "Topic Distribution"
"dataSource": {"generativeInsights": "What are the topics?"}
}
}
]
}
}
]
}
}
Curl
Run the following curl command. In this example, the payload is saved in create_dashboard_nested.json.
curl -X POST \ -H "Authorization: Bearer $(gcloud auth --impersonate-service-account=$SA_EMAIL print-access-token)" \ -H "Content-Type: application/json; charset=utf-8" \ -d @create_dashboard_nested.json "https://$ENDPOINT/v1/projects/$PROJECT/locations/$LOCATION/dashboards?dashboard_id=$DASHBOARD_ID"
Edit dashboard
Follow these steps to edit your dashboard.
Console
Choose a dashboard and click More more_vert > Edit to edit the dashboard.
Choose one of the following options:
- + Add tile
- + Add section
- edit Rename tab
- content_copy Duplicate tab
- delete Delete tab
- Choose a date range to view your results.
- Choose a tile and click More more_vert. A side panel appears with the following options: Ask AI, Tile settings, and Query editor.
Additional options
Ask AI: Click Ask AI and enter a text description of your changes. The AI assistant can help you in building your dashboard. You also have an option to ask follow-up questions to the AI assistant. If you click Save Chart, you can rollback to the previous visualizations. Natural language (NL) lets you to generate visualizations by specifying descriptive queries like What is the average conversation count each day for a virtual agent? These queries are converted into SQL for execution against BigQuery and visual specification (VEGA lite) for display purposes.
Tile settings: In this section, you can update the visual specification directly to control the chart attribute. For example: Change the display name of the chart. You can choose your preferred chart type in the drop-down menu. The chart API supports the following chart types and data representations:
- BAR
- LINE
- AREA
- PIE
- SCATTER
- TABLE
- SCORE_CARD
- SUNBURST
- GAUGE
- SANKEY
You can further specify the conversation filter or new time range upon reloading the dashboard.
Query editor: You can interact with the AI assistant and build a dashboard. If you are not satisfied with the AI-generated SQL response, you can update the SQL directly to regenerate a chart and the SQL response further repopulates the existing visual specification. The query editor displays the generated SQL, whose table name is masked to reference a predefined placeholder string, INSIGHTS_DASHBOARD_VIEW. Note: You can only do any SELECT operation.
Curl
Run the following curl command. In this example, the payload is saved in update_dashboard.json.
curl -X PATCH \ -H "Authorization: Bearer $(gcloud auth --impersonate-service-account=$SA_EMAIL print-access-token)" \ -H "Content-Type: application/json; charset=utf-8" \ -d @update_dashboard.json \ "https://$ENDPOINT/v1/projects/$PROJECT/locations/$LOCATION/dashboards/$DASHBOARD_ID?update_mask=root_container"
Delete dashboard
Follow these steps to delete a dashboard and all related charts.
Console
Choose a dashboard and click More more_vert > Delete to delete the dashboard.
Curl
Run the following curl command.
curl -X DELETE \ -H "Authorization: Bearer $(gcloud auth --impersonate-service-account=$SA_EMAIL print-access-token)" \ "https://$ENDPOINT/v1/projects/$PROJECT/locations/$LOCATION/dashboards/$DASHBOARD_ID"