Creating Dashboards¶

Build your first dashboard with Dango data in Metabase.

Overview¶

Dashboards in Metabase combine visualizations from your Dango data into interactive views. This guide covers the Dango-specific workflow.

Quick Start: First Dashboard¶

Create a revenue dashboard in 5 minutes.

Step 1: Create a Question¶

1. Open the Web UI at http://localhost:8800 and click Open Metabase in the sidebar (SSO bridge handles login automatically)

   Alternatively, access Metabase directly at http://localhost:3000. Check .dango/metabase.yml for credentials.

2. Click "+ New" → "Question"

3. Select "DuckDB" database
4. Choose table: marts.customer_metrics (or any marts table)
5. Summarize: Sum of lifetime_value
6. Group by: created → by Month
7. Visualization: Line chart
8. Click "Save" → Name: "Monthly Revenue"

Step 2: Create Dashboard¶

1. Click "+ New" → "Dashboard"
2. Name: "Revenue Dashboard"
3. Click "Create"

Step 3: Add Question to Dashboard¶

1. Click "Add a saved question"
2. Select "Monthly Revenue"
3. Resize and position
4. Click "Save"

Done! Your first dashboard is ready.

Querying Dango Data¶

Use Marts Tables¶

For dashboards, query pre-aggregated marts instead of raw data:

-- Good: Query marts (fast, clean)
SELECT * FROM marts.customer_metrics
ORDER BY lifetime_value DESC
LIMIT 10

-- Avoid: Query raw tables (slow, may have duplicates)
SELECT * FROM raw_stripe.customers

Marts are designed for business intelligence - use them.

Example Questions¶

Monthly Revenue Trend (Line chart):

SELECT
    DATE_TRUNC('month', created_at) as month,
    SUM(lifetime_value) as revenue
FROM marts.customer_metrics
GROUP BY month
ORDER BY month

Top 10 Customers (Table):

SELECT
    email,
    lifetime_value,
    lifetime_orders
FROM marts.customer_metrics
ORDER BY lifetime_value DESC
LIMIT 10

Customer Count (Number):

SELECT COUNT(*) as total_customers
FROM marts.customer_metrics

Working with dlt Metadata¶

Query raw data with dlt timestamp columns:

SELECT *
FROM raw_stripe.charges
WHERE _dlt_extracted_at > CURRENT_DATE - INTERVAL 1 DAY
ORDER BY _dlt_extracted_at DESC

Dashboard Layout¶

Recommended Structure¶

Executive Dashboard:

[  KPI 1  ] [  KPI 2  ] [  KPI 3  ]
[        Trend Chart (Line)       ]
[ Bar Chart ] [ Table/Details ]

Key principles:

• Most important metric at top-left
• Trends over time below KPIs
• Details at bottom

Adding Filters¶

1. Click "Add a filter" (top-right)
2. Choose Time → Select date field
3. Connect to cards with date columns
4. Set default: "Last 30 days"

Date filters work across all cards automatically.

Data Refresh¶

Dashboards show live data from DuckDB:

# 1. Sync new data
dango sync

# 2. Run transformations
dango run

# 3. Reload dashboard (data updates automatically)

No additional steps needed - Metabase queries live data.

Best Practices¶

1. Pre-aggregate in dbt¶

Create dbt marts for dashboard queries:

-- dbt/models/marts/revenue_by_month.sql
{{ config(materialized='table') }}

SELECT
    DATE_TRUNC('month', created) as month,
    SUM(amount) as revenue,
    COUNT(*) as order_count
FROM {{ ref('stg_stripe_charges') }}
WHERE status = 'succeeded'
GROUP BY month

Then query the mart in Metabase - much faster.

2. Keep Dashboards Focused¶

• One dashboard = one purpose
• 6-12 cards maximum
• Use filters instead of multiple similar dashboards

3. Name Questions Clearly¶

Good: "Monthly Revenue by Product Category" Bad: "Analysis 1" or "test query"

4. Version-Control Your Dashboards¶

Build dashboards locally, not directly on your cloud Metabase instance:

1. Build locally — create and iterate on dashboards at http://localhost:3000
2. Export — run dango metabase save to export dashboards as YAML to your project repo
3. Commit — git add the saved YAML files and commit. Dashboards are now version-controlled
4. Push — dango remote push deploys your project to the cloud server. Transfer the metabase/ directory separately (e.g., scp -r metabase/ root@server:/srv/dango/project/)
5. Restore — run dango metabase load on the cloud server to import dashboards from the transferred files

Why this matters:

• A cloud server crash or rebuild loses all Metabase state (saved questions, dashboards, collections)
• Backups cover Metabase, but the git repo is the authoritative source of truth
• Treat dashboards like code — review changes, track history, roll back if needed

If you must build directly on cloud:

• Run dango metabase save immediately after making changes on the cloud instance
• Pull the saved YAML files back to your local repo and commit them

See Save & Load for full details on the export/import workflow.

Troubleshooting¶

Dashboard Not Loading¶

Check if dbt models ran successfully:

dango run
duckdb data/warehouse.duckdb "SELECT COUNT(*) FROM marts.customer_metrics"

Missing Tables in Metabase¶

Sync the database schema:

1. Admin → Databases → DuckDB
2. Click "Sync database schema now"

Slow Dashboards¶

1. Query marts instead of raw tables
2. Add date filters to limit data
3. Create more specific dbt models

Next Steps¶