Dashboards

Dashboards let you describe what you want to track in plain English and get a fully functional analytics interface in seconds. The AI generates optimized queries, selects appropriate visualizations, and configures interactive components—all from your Supabase data.

Prerequisites

You need at least one connected Supabase project in your Dreambase workspace. The dashboard will query your project's database to generate visualizations.

Creating a Dashboard

Step 1: Start a New Dashboard

From your Workspace Overview or the Dashboards page, click New Dashboard. You'll see a prompt input where you describe what you want to track.

Step 2: Describe What You Want

Type a natural language description of the metrics and data you want to see. Be specific about:

• Metrics: "signups", "revenue", "active users"
• Time ranges: "this month", "last 7 days", "YTD"
• Breakdowns: "by plan", "by region", "by feature"
• Comparisons: "vs last month", "week over week"

Example prompts:

• "Show me daily active users, revenue by plan, and our top 10 features this month"
• "User signups over time with conversion rates and churn"
• "Error rates, response times, and system health for the past week"

Step 3: Wait for Generation

The AI processes your prompt and:

1. Analyzes your database schema
2. Generates optimized SQL queries
3. Selects appropriate chart types
4. Configures metrics, tables, and filters
5. Renders the dashboard

This takes a few seconds. You'll see a building indicator while the dashboard generates.

Step 4: Interact with Your Dashboard

Once generated, your dashboard is fully interactive:

• Filter data using date pickers and dropdowns
• Hover over charts for detailed tooltips
• Click chart elements to drill down
• Switch chart types within visualization families

Dashboard Components

Metrics (KPI Cards)

Large, scannable numbers for headline metrics. Each metric card shows:

• Current value with formatting (currency, percentage, number)
• Comparison indicator vs previous period (optional)
• Trend direction with color coding

Charts

Interactive ECharts visualizations across five families:

Switching chart types: Within the same family, you can switch chart types instantly (e.g., bar to line) without regenerating queries. Cross-family switches require regeneration.

Tables

Sortable, paginated data views for detailed exploration:

• Click column headers to sort
• Paginate through large datasets
• Click rows to view detail pages (if configured)
• Column widths and formatting are auto-configured

Lists

Ranked items for leaderboards and top-N views:

• Primary label, secondary label, and value
• Icon mappings for categories
• Customizable sorting

Filters

Cross-filter all components simultaneously:

• Date range: Select custom date ranges
• Select: Single-value dropdown filters
• Multi-select: Multiple-value filters

Filters apply client-side for instant response. All charts, tables, and metrics update together.

Multi-Dataset Architecture

Dashboards use 2-4 purpose-built datasets optimized for different component types:

This separation ensures fast rendering. Each query is optimized for its specific job rather than trying to power everything from one bloated query.

Refining Dashboards with Chat

Dashboards rarely perfect on the first try. Use chat to iterate:

1. Click the Chat button on your dashboard
2. Describe what you want to change
3. The AI updates queries and components in place

Example refinements:

• "Add a breakdown by region"
• "Change the bar chart to show percentages"
• "Filter to only paid users"
• "Add a table showing individual transactions"
• "Remove the funnel chart"

Data Refresh

Dashboards cache query results for fast loading. Refresh data two ways:

Manual Refresh

Click the Refresh button to re-run all queries and update the cache. Use this when you need the latest data immediately.

Scheduled Refresh

Configure automatic refresh on a schedule:

1. Open dashboard settings
2. Enable Scheduled Refresh
3. Set the interval (default: 30 minutes)
4. Choose your timezone

Scheduled refreshes run in the background. The dashboard shows the last refresh timestamp.

Data Storage & Security

Your Data Stays Yours

Dashboard data is cached in your own Supabase project:

• Results stored in a dreambase storage bucket
• Compressed as gzip JSON for efficiency
• Never leaves your infrastructure

Dreambase stores only:

• Query definitions (SQL)
• Component configurations
• Layout and settings

Query Safety

All queries are validated before execution:

• EXPLAIN validation prevents expensive queries
• Row limits cap result sizes
• Timeouts prevent runaway execution
• Formula guardrails block code injection

Dashboard Lifecycle

Failed dashboards can be regenerated or refined via chat to fix issues.

Tips for Better Dashboards

Be Specific in Prompts

Instead of: "Show me user data"

Try: "Show me daily active users, signup conversion rate, and top 5 features by usage this month"

Start Simple, Then Refine

Generate a basic dashboard first, then use chat to add components. It's faster than trying to specify everything upfront.

Use Filters Strategically

Add filters for dimensions you'll frequently slice by: date ranges, user segments, regions, plans.

Match Charts to Data

• Time series → Line or area charts
• Comparisons → Bar charts
• Proportions → Pie or donut
• Rankings → Tables or lists

Troubleshooting

Dashboard Won't Generate

• Check your Supabase connection is active
• Verify your prompt references tables that exist
• Try a simpler prompt first

Data Looks Wrong

• Check filter settings
• Verify the date range
• Click Refresh to get latest data

Charts Are Slow

• Large datasets may need pagination
• Consider adding filters to reduce data volume
• Check if scheduled refresh is running

Next Steps

• Creating Projects - Set up your Supabase connection
• Sharing Reports - Share dashboards with your team
• Workspace Settings - Configure refresh schedules