Relations tell dScribe how different assets in your catalog connect to each other — that a report uses a particular dataset, that a definition is mentioned in a dashboard, that a metric is related to another one. They're the backbone of the catalog: they power impact analysis, navigation between assets, contextualized search, and data lineage.
You'll find an asset's relations on the Relations tab of its detail page. From there you can view, edit, and explore them in three different ways: a table view, a graph view, and a hierarchy graph view.
How relations get created
Some relations are added by hand, others appear automatically. There are three sources:
Automatic lineage from crawled sources
When dScribe ingests assets from a source system (Power BI, Tableau, Snowflake, BigQuery, …), it also detects how those assets relate to each other and creates the matching relations automatically.
For example, when you connect a Power BI workspace, dScribe registers the relations between the workspace, the reports it contains, and the semantic models those reports build on — so you get cross-source lineage without having to draw a single line by hand.
[Screenshot: hierarchy graph of a Power BI report showing automatic lineage to its semantic model and upstream definitions]
From @-mentions
When you type @ in any rich-text field (most often the Documentation field) and pick an asset from the suggestions, dScribe creates a mentions / is mentioned in relation between the two assets. The mention stays clickable in your text, and the relation shows up on the Relations tab of both assets.
Manually on the Relations tab
Editors can add extra relations directly from the Table view on the Relations tab. Click Add relations, pick the relation type and the asset to link to, and (optionally) add a short description explaining the relationship. To remove a relation, click the disconnect icon at the end of the row.
The three views on the Relations tab
All three views show the same relations — they just present them differently. Switch between them with the toggle at the top of the Relations tab: Table, Graph, or Hierarchy graph.
Different default per asset type: Administrators can choose which of the three views is the default on each asset type. A dataset might open in the hierarchy graph (where lineage matters most), while a definition might open in the graph view. See Asset Types for the config.
Table view
The table view lists every related asset on a single row, with columns for the Relation Type, the related asset's Name, its Asset Type, and an optional Relationship description. This is the view to use when you want to edit — add relations, remove them, or annotate them with a short description of the link.
Graph view
The graph view renders relations as a network centered on the current asset. Each node is an asset, color-coded by type (Definition, Report, Dataset, …), and each edge is labeled with its relation type.
The graph is interactive:
Click any node to open a preview panel on the right showing that asset's documentation and key info.
From the highlighted node, click the book icon to jump to that asset's full detail page.
Click the graph icon on a node to expand all of that asset's relations into the graph.
A legend on the left shows which colors map to which asset types; a filter on the right lets you narrow the graph by asset type or relation type so you can focus on what matters.
Hierarchy graph view
The hierarchy graph is built for lineage. It arranges assets left-to-right based on data flow — upstream sources on one side, downstream consumers on the other — and highlights the data lineage relations between them.
Two controls help you tune the view:
Lineage depth — how many degrees of upstream and downstream dependencies to show. Increase it to see further across the lineage chain; decrease it to keep the picture focused.
Suggest relationships — let AI propose additional relations that might be missing (see next section).
Letting AI suggest relations
Click Suggest relationships in the hierarchy graph and dScribe's AI scans the asset, its documentation, and the surrounding catalog to find connections that aren't yet captured as relations. Each suggestion is added to the graph as a dashed line with a lightbulb icon.
Click a suggestion to open the Relation Details panel, which shows the proposed relation type, the two assets being connected, and an optional description. From there you can:
Accept — the suggestion becomes a regular relation and appears in all three views.
Reject — the suggestion is dismissed.
Custom relation types
dScribe ships with a set of standard relation types that cover the most common needs:
is used in / uses
is related to
is parent of / is child of
mentions / is mentioned in
is a duplicate of / has a duplicate
If you need something specific to your organization — certifies, governs, applies to, anything — administrators can create custom relation types from the admin portal. Once created, the new type becomes available everywhere relations can be added.
Where to go next
→ Editing assets — how to edit relations alongside the rest of an asset's documentation
→ Asset Types — set the default relations view per asset type
Have a question or can't find what you're looking for? Use the chat icon inside the catalog to reach the dScribe support team.