Users need the Admin role to create and edit tiles. Analysts have read-only access. Tile principles
-
Every tile has two parts: The data (SQL, Python, or markdown) and the visualization (YAML configuration). You define what to show, then how to show it.
-
Fi handles most of the work: Describe what you want in plain English, and Fi generates both the query and visualization. You don’t need to write SQL or configure YAML unless you want to.
-
Tiles are composable: Start simple with a few KPIs, then layer in charts and tables. Each tile is independent, so you can mix and match freely.
Tile types
Definite supports four tile types:
| Type | Purpose | Best for |
|---|
| KPI | Display single metric values prominently | Revenue totals, counts, percentages |
| Chart | Visualize data graphically | Trends, comparisons, distributions |
| Table | Show data in rows and columns | Detailed data, drill-downs |
| Markdown | Add formatted text | Headers, descriptions, documentation |
When should I use each tile type?
| If you want to… | Use |
|---|
| Highlight a single number executives check daily | KPI |
| Show how a metric changes over time | Chart (line or area) |
| Compare values across categories | Chart (bar) |
| Show proportions of a whole | Chart (pie) |
| Let users explore raw data or drill into details | Table |
| Add context, headers, or documentation | Markdown |
Creating tiles
Using Fi (recommended)
The fastest way to create tiles is with Fi. Just describe what you want:
- “Show me total revenue collected in the last 90 days as a KPI”
- “Create a line chart of monthly signups for the past year”
- “Add a table of our top 10 customers by lifetime value”
Fi generates the SQL query, picks the right visualization, and configures the tile. You can refine it with follow-up prompts like “change the color to green” or “add a comparison to last period.”
Even technical users find Fi faster than writing queries from scratch. Start with Fi, then customize the generated code if needed.
Creating tiles manually
If you prefer to write your own queries:
- Click the + button in the Tiles panel on the left sidebar
- Choose how you’ll define the data:
- SQL: Query your connected data sources
- Python: Run Python code for complex transformations
- Markdown: Add formatted text for headers and documentation
- Write your query or content
- Configure the visualization using the settings panel or YAML
Tile configuration with YAML
Every tile’s visualization is controlled by YAML. Here’s an example of a KPI tile showing payment success rate:
# Define the data source
engine: "sql"
sql: |
SELECT
ROUND(
(COUNT(CASE WHEN status = 'succeeded' THEN 1 END)::DECIMAL / COUNT(*)::DECIMAL),
4
) as success_rate_percent
FROM LAKE.STRIPE.charges
WHERE to_timestamp(created) >= CURRENT_DATE - INTERVAL '30 days'
# Configure the visualization
viz:
type: "kpi" # Tile type: kpi, chart, table, or markdown
series:
- field: "success_rate_percent"
format: "percent" # Display as percentage
colors:
color: "#10b981" # Green color for the metric
style:
showLongNumber: false # Use compact number formatting
engine specifies SQL or Python. The sql (or python) block contains your query. The viz block controls how it displays: type, formatting, colors, and styling.
Common tile properties
All tiles share these properties:
| Property | Description |
|---|
id | Unique identifier for the tile |
type | The tile type (kpi, chart, table, markdown) |
title | Optional title displayed above the tile |
x, y | Grid position (column, row) |
w, h | Size (width and height in grid units) |
source | Reference to a dataset (for kpi, chart, table) |
Layout and organization
Tiles are arranged on a grid system. Click the Edit button in the top toolbar to enter edit mode, then:
- Drag tiles to reposition them
- Resize tiles by dragging their edges
- Group related tiles together for better readability
Start with KPIs at the top for high-level metrics, add charts in the middle for trends, and include tables at the bottom for detailed data.
Next steps
