Superset Dashboard

Step 2: Register Dataset (create_dataset)

Register a table or SQL query as a Superset dataset.

Physical dataset (table created by write_query or already present in a Superset-connected BI database):

create_dataset(name="materialized_table_name", database_id="<from step 1 or list_bi_databases>")

Virtual dataset (aggregated/transformed view):

create_dataset(name="view_name", database_id="<from step 1 or list_bi_databases>", sql="SELECT ... FROM materialized_table_name")

• Returns dataset_id — save it for Step 3.
• IMPORTANT: create_dataset.database_id must be a Superset BI database connection ID from write_query or list_bi_databases(). Do NOT assume a source connector ID is valid here.

Step 3: Create Charts (create_chart)

Create visualization charts referencing the dataset.

create_chart(
    chart_type="bar",        # bar, line, pie, table, big_number, scatter
    title="Chart Title",
    dataset_id="<from step 2>",
    metrics="revenue,COUNT(order_id)",
    x_axis="date_column",
    dimensions="category"
)

Metrics format:

• Plain column name defaults to SUM(column): "revenue" -> SUM(revenue)
• Explicit aggregation: "AVG(price)", "MAX(amount)", "MIN(cost)", "COUNT(id)"
• Multiple metrics comma-separated: "revenue,COUNT(order_id),AVG(price)"

For big_number charts:

• Use a single metric: metrics="AVG(activity_count)"
• No x_axis or dimensions needed.

Step 4: Create Dashboard (create_dashboard)

create_dashboard(title="Dashboard Title", description="Optional description")

Returns dashboard_id — save it for Step 5.

Step 5: Add Charts to Dashboard (add_chart_to_dashboard)

add_chart_to_dashboard(chart_id="<from step 3>", dashboard_id="<from step 4>")

Repeat for each chart.

Step 6: Validate the Published Dashboard (bi-validation)

After assembling the dashboard, load bi-validation and verify:

• the dashboard exists and is reachable via get_dashboard
• the expected charts appear via list_charts
• each chart is inspected via get_chart
• chart titles, types, metrics, x-axis fields, dimensions, and dataset wiring match the intended configuration
• each chart is checked with get_chart_data to confirm it runs without backend errors
• key numeric values are compared where expected results or tolerances are available

Do not report success until this validation pass completes.

Viewing & Querying

Updating

• Update dashboard: update_dashboard(dashboard_id, title="New Title", description="New desc")
• Update chart: update_chart(chart_id, title="...", chart_type="...", metrics="...", x_axis="...", description="...")
• If the wrong dimensions or dataset_id was used, recreate the chart instead of relying on update_chart.

Deleting

MUST confirm with user before any deletion.

• delete_dashboard(dashboard_id="...")
• delete_chart(chart_id="...")
• delete_dataset(dataset_id="...")

Important Rules

1. Use write_query only when needed to materialize source-side results into a Superset-accessible table.
2. Use a Superset BI database ID for create_dataset — obtain it from write_query or list_bi_databases, not from a source connector.
3. Choose the correct chain:
   • materialized flow: write_query -> database_id -> create_dataset -> dataset_id -> create_chart
   • direct BI flow: list_bi_databases -> create_dataset(... or sql=...) -> dataset_id -> create_chart
4. Language: Match the user's language (Chinese input -> Chinese output).
5. Multiple charts: Create separate datasets for charts needing different data shapes.