6. Data Overview ui-expert-blue

DISQOVER aggregates data from disparate data sources into searchable instances. Multiple sources often contribute to a single instance, each defining a subset of the instance’s properties. The provenance of this information is a key aspect of DISQOVER and is prominently featured in the dashboard environment.

The data sources overview is a separate environment dedicated solely to the data sources and their links, that helps to gain additional insights. In this environment, you can quickly explore data sources, see which canonical types each source contributed to and examine the links between canonical types.

Information about the available data sources and configuration is available under the data (ui-data) tab in the top menu bar. Within the tab, you can choose a table view (“Data sources”) or a visual representation (“Ontology graph”) of this information by clicking the respective title at the top of the page. To see an overview of the canonical types with their properties and facets, click “Configuration”.

6.1. Data sources ui-expert-blue

The tabular overview of the data sources contains three columns (Figure 6.1):

  1. Title: the name of the data source
  2. Date: the day of the last update
  3. #Instances: the number of instances to which this data source contributes
image

Figure 6.1 Overview of the available data sources.

If you want to review a data source in more detail, you click the entry of that source. This opens the source detail page (Figure 6.2). You could also access this page through the instance details in the dashboard, by selecting a source and clicking the icon (ui-externallink) as described in section 3.2.2.

At the top of this page, you can see the name of the data source. Below the title, you can see how many instances in which canonical types contain information contributed by this source. If you click one of these tiles, you open a new dashboard containing the instances of the selected canonical type that contain information originating from this data source.

Below the overview of canonical types, there is a description of the data source, the date of the last update, a link to its homepage and the URI of the datasource in DISQOVER. The “Scheme URI” is the template used to build the URIs of instances defined by the data source. The scheme example URI is an example of a possible instance URI.

image

Figure 6.2 Detail of one of the available data sources in DISQOVER: PubMed.

6.2. Ontology graph ui-expert-blue

The ontology graph shows a visual summary of the linkage structure of the data in DISQOVER. The canonical types are represented as nodes and the links between canonical types as lines.

6.2.1. Network representation ui-expert-blue

The ontology graph page (Figure 6.3) is divided into two parts: a list of data sources and graph representation.

Canonical types, as represented by their icons, are the nodes of the ontology graph. The size of a node is proportional to the number of instances in that canonical type. The number of instances is also shown within each node.

Nodes of different canonical types are connected with a line (or “edge”) if there is a link between at least one pair of instances of these canonical types. Similarly, a self-linking edge represents links between instances from the same canonical type.

In the top right corner of the graph, you can choose a network layout:

  • Autofocus (default): the network representation is built and optimized around a selected canonical type [1].
  • Automatic: the network representation is built and optimized for the entire ontology graph.
  • Manual: no background optimization of the network representation. The graph becomes static, and you can manually drag canonical types to adjust the representation.
image

Figure 6.3 Overview of the ontology graph.

6.2.2. Data source details ui-expert-blue

By default, all sources are included in the ontology graph. You can check or uncheck data sources in the list. The text search field helps you search for data sources by name.

When you check a data source, you notice (Figure 6.4):

  • The canonical types to which the data source contributes are highlighted.
  • The blue links are links created by this data source. The other links are faded and colored yellow.
image

Figure 6.4 Network representation of the ontology graph when PubMed is selected.

6.2.3. Canonical type details ui-expert-blue

Network highlighting

Clicking a canonical type in the network representation highlights part of the graph (Figure 6.5). A lock icon (ui-lock) appears in the top right corner of the screen, next to the canonical type name. The following visual changes occur:

  • The clicked node is highlighted. The thickness of the node’s surrounding border increases.
  • All canonical types that have a direct link with the clicked canonical type (first-degree relations) are annotated with the number ‘1’. The thickness of the linking lines increases.
  • Canonical types that share a link with these first degree relations but not with the clicked canonical type (second-degree relations), are annotated with the number ‘2’.

To release the clicked node and remove the highlighting, you click the lock (ui-lock).

image

Figure 6.5 Network representation of the ontology graph with the highlighted canonical type of interest and its first- and second- degree relations.

Canonical type pop-up

When clicking a canonical type in the ontology graph, a window pops up providing the following information about that canonical type (Figure 6.6):

  • Contributing data sources: these are the data sources that contribute to the instances of this canonical type. You can check or uncheck these sources and see the changes to the network as described in the previous section.
  • Links to: A list of all first neighbors of this canonical type. For each canonical type in this list, you see all data sources that establish a link between that canonical type and the clicked canonical type. Icons show how the links are used:
    • to define properties shown in the instance details (ui-arrowright)
    • to define facets that can be used in widgets (ui-filter)
    • to define to define typed links (ui-link)
image

Figure 6.6 Overview of data sources contributing to the “Clinical Study” canonical type and the links to other canonical types.

6.3. Configuration ui-expert-blue

The configuration overview (Figure 6.7) shows the configuration of canonical types and typed links. By clicking the arrows (ui-chevron-right) next to the headers, you can expand or collapse sections.

Firstly, the configuration overview contains a list of all canonical types, their URIs (when clicking ui-uri) and their content:

  • Facets: their name, description and data type. By clicking the URI-icon (ui-uri) next to their name, you see the URI and other configuration information.
  • Properties: their name, description and data type. By clicking the URI-icon (ui-uri) next to their name, you see the URI and other configuration information.
  • Subinstances: by clicking the URI-icon (ui-uri) next to their name, you see the URI and other configuration information.

Secondly, you see a list of all typed links between canonical types. When they are expanded, you see which relation types have been configured, with their URI and configuration information available after clicking the URI-icon (ui-uri).

The provenance of the information is easily visualized. On the left side of the screen, you see a list of data sources. By default, they are all included in the configuration overview, but you can check or uncheck individual datasources as well. If you check a datasource, the configuration overview updates to show only the contribution of that datasource.

image

Figure 6.7 Overview of the DISQOVER configuration, showing for example the facets for the Active Substance canonical type.

[1]The motion of the nodes in the network representation of the ontology graph is temporary and the result of the ongoing optimization of the representation in the background.