Visual Discovery Graph overview

The Visual Discovery Graph is a powerful feature in the Witboost Marketplace that allows users to explore relationships between systems in a visual and interactive way.
It provides an intuitive representation of key metadata dimensions, helping users to discover connections across the catalog.

Built-in heatmap layers

By default, the Visual Discovery Graph includes several predefined heatmap layers that provide insights into the structure and governance of systems. These layers are available without additional configuration:

Label	Kind	Key	Description
Connections	number	CONNECTIONS	Sum of input and output systems connected to the system.
Violations	number	RED_FLAGS	Number of governance violations detected on the system.
Output Ports	number	COMPONENTS	Number of consumable components exposed by the system.
Reviews	number	REVIEWS	Number of reviews associated with the system.
Publication Date	date	PUBLISHED_AT	Date the system was published.

These built-in layers are computed dynamically based on internal system metadata and can be used alongside custom layers defined via configuration.

Configure custom heatmap layers

In addition to the built-in layers, organizations can configure custom heatmaps to reflect business-specific dimensions of the data catalog.

Custom layers are defined under the mesh.marketplace.ui.visualDiscoveryGraph.layers property as an array of objects, where each object describes a single layer.

Here’s an example of how to configure them:

mesh:
  marketplace:
    ui:
      visualDiscoveryGraph:
        layers:
          - label: 'kind'
            kind: list
            key: 'KIND'
            path: 'kind'
          - label: 'Owner'
            kind: list
            key: 'OWNER'
            path: 'dataProductOwnerDisplayName'
          - label: 'Taxonomy'
            kind: list
            key: 'TAXONOMY'
            path: 'taxonomy.name'

Layer properties

Each custom layer supports the following properties:

• label: The name shown in the UI for this layer.
• kind: The type of visualization. Allowed values are:
  • list: for categorical values (e.g., kind, taxonomy, owner)
  • number: for numeric values (e.g., count, size)
  • date: for temporal values (e.g., publication date)
• key: A unique identifier for the layer. Should be written in uppercase.
• path: The field extracted from the system descriptor

---

How to add a new custom layer

To add a new layer:

1. Identify the field in the system descriptor you want to use as the data source.
2. Choose the appropriate kind for how the value should be visualized:
   • Use list for distinct string values.
   • Use number for numeric data.
   • Use date for timestamps.
3. Define a unique key for the layer (uppercase recommended).
4. Set the path to the correct metadata field using dot notation.

---