Skip to content

Catalog Page

Browse the data catalog with search, profiling, lineage, and impact analysis.

Overview

The Catalog page (/catalog) is a searchable data catalog that lets you explore all sources and dbt models in your project. View column profiles, PII annotations, schema drift, dependency lineage, and downstream impact — all from the browser.

Navigate to Catalog in the top navigation bar.

The Catalog has three views: List (default), Detail, and Lineage. Your position is reflected in the URL, so you can bookmark or share specific views.

Summary Cards

At the top of the list view, four cards show high-level counts:

Card Description
Sources Number of data sources
Tables Total table count across all sources
Models Number of dbt models
Sources Fresh Count of fresh sources vs. total (green if all fresh, amber if some stale)

Source Breakdown

Below the summary cards, a grid of per-source cards shows each source with:

  • Source name and table count
  • Estimated row count
  • Freshness badge — "fresh" (green), "stale" (amber), or "unknown" (gray)

Click a source card to filter the table below to only that source's tables. Click Show X more to reveal additional sources beyond the initial 6.

Type Filter Tabs

Filter the model/source list by type:

Tab Shows
All Everything
Sources Raw source tables (dlt-ingested data)
Staging stg_ staging models
Intermediate int_ intermediate models
Marts Final analytics models

Each tab shows a count of matching items. When a source filter is active, a "Showing tables from source: X" indicator appears with a Show all link to clear the filter.

The search bar (top of the list view) searches across model names, column names, and descriptions with a 300ms debounce. Results show:

Column Description
Name Model or source name (clickable to open detail)
Type Type badge (source, staging, intermediate, marts)
Match What matched — "name", "description", or "column" with the matched column name
Description Truncated model description

Click Clear next to the results header to return to the filtered list.

Model / Source List

The main table displays all items matching the current tab and source filters:

Column Description
Name Model or source name (clickable to open detail view)
Type Color-coded type badge
Description Truncated model description
Tests Pass/fail count (e.g., "3 pass / 1 fail")
Documented Column documentation ratio (e.g., "⅝")
Tags Up to 3 tag badges, with "+N" for additional tags

Click any row to open the Detail View.

Detail View

Clicking a model or source opens the detail view with a breadcrumb ("Catalog / model_name") and three sub-tabs.

  • Model name and type badge
  • Description (if available)
  • Metadata row: materialization type, row count, and profiling timestamp ("Profiled 5m ago")
  • Refresh Profile button (source tables only) — re-runs column profiling

Columns Tab

A table of all columns in the model:

Column Description
Column Name Column name with optional PII and drift badges (see below)
Type Data type (e.g., VARCHAR, INTEGER, TIMESTAMP)
Description Column description from dbt schema YAML
Nullable Yes/No
Tests Test badges — green "✓" for passing, red "✗" for failing
Null % Percentage of null values
Distinct Count Number of distinct values
Min Minimum value
Max Maximum value

PII Badges

Columns detected as containing personally identifiable information show badges:

  • PII ⚑ (amber) — auto-detected PII, with entity type in tooltip
  • PII ✓ (green) — confirmed PII annotation

Schema Drift Badges

Columns with detected schema drift show an orange badge with the drift type (e.g., "type_changed", "column_added").

Code Tab

Shows the model's SQL (not available for source tables):

  • Raw SQL (with Jinja) — the source .sql file content
  • Compiled SQL — the rendered SQL after Jinja processing

SQL is syntax-highlighted for readability.

Info Tab

Additional model metadata:

  • Description — full model description
  • Schema, Materialization, Type — key properties
  • Tags — tag badges
  • Tests — list of tests with status dots (green/red/gray)
  • Upstream Dependencies — clickable links to models this one depends on
  • Downstream Dependents — clickable links to models that depend on this one
  • View in Lineage button — opens the lineage view focused on this model

Lineage Visualization

Click the Lineage toggle button (top-right of the list view) to switch to the DAG (directed acyclic graph) view showing data flow from sources through staging, intermediate, and marts models.

  • Zoom in/out — use the + and − buttons, or scroll
  • Fit all — click the ↔ button to fit the entire graph in view
  • Pan — click and drag the background
  • Click a node — opens the node detail sidebar

Focus Mode

From the detail view, click View in Lineage to focus the graph on that model. In focus mode:

  • A Depth selector dropdown controls how many hops to show: 1, 2, 3, or 5
  • A Show Full DAG button returns to the complete graph
  • The URL updates to reflect the focused node

Node Detail Sidebar

Clicking a node in the graph opens a sidebar with:

  • Type and Materialization
  • Description
  • Tests count
  • Doc Coverage — documented columns vs. total

Actions:

  • View Details — opens the full detail view for this model
  • Show Downstream — highlights all downstream dependents in the graph
  • After highlighting, a count shows "X downstream model(s)" with a Clear Highlight button
How it works
  • List models: GET /api/catalog/models returns all models and sources with overview counts, freshness data, and per-source detail.
  • Model detail: GET /api/catalog/models/{name} returns columns with profiling stats, tests, raw/compiled SQL, dependencies, and dependents.
  • Refresh profile: POST /api/catalog/{source}/{table}/profile re-runs column profiling for a source table.
  • Search: GET /api/catalog/search?q={query} searches across model names, descriptions, and column names (2-100 character queries).
  • Lineage: GET /api/catalog/lineage returns nodes and edges for the full DAG.
  • Impact analysis: GET /api/catalog/impact/{model_name} returns the downstream dependency tree with total count.
  • PII and drift data are fetched in parallel on page load and merged onto column data in the detail view.
  • URL state: The page encodes view, model, tab, focus, source filter, and type filter into URL parameters for bookmarkability and browser back/forward support.

Troubleshooting

"No models found" message
dbt has not been run yet. Run dbt run from the CLI to materialize models and generate the manifest, or trigger a sync which runs dbt automatically.
Column profiling stats show "—"
Profiling has not been run for this table. Click Refresh Profile (available on source tables) or wait for the next scheduled sync which includes profiling.
Lineage graph is empty
The lineage graph requires a dbt manifest with dependency information. Run dbt run or dbt compile to generate the manifest.