Cluster visualization

Cluster visualization renders your cluster data as an interactive map allowing you to see a quick overview of your cluster sets and quickly drill into each cluster set to view subclusters and conceptually-related clusters.

This can assist you with the following actions:

This page contains the following sections:

See these related pages:

Visualizing a cluster

Cluster visualization has now been integrated with your Documents tab and you can add a cluster visualization widget directly to your Dashboard.

To view multiple cluster sets, you must create different dashboards. There can be only one cluster visualization widget on a dashboard at a time.

    Notes:
  • To use Cluster Visualization, you must have workspace permissions to the clusters browser. See Browsers in the Workspace permissions section.
  • Permissions for Cluster Set object and the multiple choice fields that hold cluster data should be kept in sync (Relativity 9.5.196.102 and above). Refer to the Cluster Set object description in Workspace permissions. If a group has the None permission set on the Cluster Set object, but View/Edit/Add/Delete permissions to the Multiple Choice fields that hold cluster data, users in that group will still see Analytics clusters in the cluster browser and will be able to visualize clusters. To prevent the group from viewing the clusters in the cluster browser (or visualizing clusters), you must item-level secure the multiple choice fields that hold the cluster data for the group.

Visualizing a cluster from the Cluster browser

To visualize a cluster from the Cluster browser, complete the following steps in your workspace:

  1. Click Cluster icon on the Documents tab to open the Cluster browser.
  2. Select a cluster on the Cluster browser.
  3. Click Visualize Cluster, or right-click the cluster and click Visualize Cluster.
  4. (Click to expand)

    Visualize Cluster button in Cluster browser
    The cluster visualization widget displays on your Dashboard and defaults to the first level of clusters that exist under the cluster you selected . If you selected a subcluster, it still defaults to the first level of clusters under the parent cluster that contains your selection.

Note: If you select a new first level cluster from the Cluster browser, the widget refreshes and the title of the widget is updated to reflect your new cluster selection.  If you select a new subcluster, the widget refreshes to display the relevant heat mapping. With any change in cluster selection in the Cluster browser, any cluster filters that have been created from selecting clusters are discarded from the search panel.

Adding a cluster visualization widget to your Dashboard

To add a cluster visualization widget to your Dashboard, complete the following steps in your workspace:

  1. Navigate to the Documents tab.
  2. Note: You do not have to be on the Clusters browser to add a cluster visualization. You can create a cluster visualization from the Folders and Field Tree browsers as well.

  1. Click Add Widget to display a drop-down menu.
  2. Select Cluster Visualization from the Add Widget drop-down menu.
  3. The Cluster Visualization Settings pop-up displays.

  1. Select the Cluster Set you want to create an interactive visual map for, and then click Apply.
  2. The cluster visualization widget displays on your Dashboard and defaults to the first level of clusters that exist under the cluster you selected .

Note: If you select a new first level cluster from the Cluster browser, the widget refreshes and the title of the widget is updated to reflect your new cluster selection.  If you select a new subcluster, the widget refreshes to display the relevant heat mapping.

Selecting a new cluster set to visualize

You can change the existing cluster set that you are visualizing in one of the following ways:

  • If you want to visualize other clusters from the Clusters browser, simply select a new cluster set in the tree on the left. The existing Cluster Visualization is updated with your new selection.
  • Note: If you select a new first level cluster from the Cluster browser, the widget refreshes and the title of the widget is updated to reflect your new cluster selection.  If you select a new subcluster, the widget refreshes to display the relevant heat mapping. With any change in cluster selection in the Cluster browser, any cluster filters that have been created from selecting clusters are discarded from the search panel.

  • From any browser, you can click the icon in the top right corner of the widget, and then select Edit from the Properties menu to launch the Cluster Visualization Settings pop-up, and then select a new cluster set.

Note: To view multiple cluster set visualizations, you must create a different dashboard for each one. There can be only one cluster visualization widget on a Dashboard at one time.

Understanding the types of cluster visualizations

You can click and pan on cluster visualizations to move around the cluster, dial or circle pack visualization at any level of depth using any of the following cluster visualizations:

All types of visualizations display a legend that is used for heat mapping. See Understanding the Cluster Visualization heat map for more information on how to use this legend.

Additionally, when you are navigating outside the Cluster browser (for example, you click on a specific Folder or field tree item), the document list contains documents that are not necessarily in the cluster set you selected. A Document Breakdown pie chart is present for both types of visualizations. This chart provides a visual breakdown of documents in the document list below (percentage of listed documents that are found in the visualized cluster set and percentage of listed documents that are not in the visualized cluster set). See Using the Document Breakdown chart.

Circle pack

Dial visualization

Nearby clusters

Using the Document Breakdown chart

The Document Breakdown pie chart provides a visual breakdown of documents in the document list below (percentage of listed documents that are found in the visualized cluster set and percentage of listed documents that are not in the visualized cluster set).

You can click on the Document Breakdown pie chart to filter the Document List on your selection (e.g., you may want to drill in and only view the list of documents that are included in the cluster set). The pie chart will display 100% after clicking on a pie section. Additionally, a filter condition is set in theSearch panel corresponding to your choice. To remove the filtering you just created when you clicked on a section of the pie chart, clear the filter condition from the search panel.

To collapse the Document Breakdown chart, click the icon. To expand the Document Breakdown chart, click the icon.

Note: The Document Breakdown is most helpful when you are navigating outside the Cluster browser on the Documents tab (e.g., folder or field browsers) and viewing documents that are not necessarily within the cluster being visualized in the dashboard widget.

Applying filters to visualized clusters

Use the search panel on the Documents tab to apply filters to your data set based on field values, saved searches, and selections made on the circle pack, dial and nearby clusters visualization panels. The filters you apply determine what documents are listed in the document list below the visualization panels and automatically apply a heat map to the circle pack, dial and nearby clusters visualization panels. See Understanding the Cluster Visualization heat map for more information regarding heat mapping.

Applying filters helps improve your review workflow and complete tasks such as the following:

  • Identify clusters containing documents with matching field values, i.e., matching coding values.
  • Mass edit and tag a list of documents created based on your visualization filters.
  • Hone in on specific clusters to examine and tag documents.
  • Perform a QC review by finding and examining clusters with a low percentage of documents coded responsive that are conceptually close to a cluster you identified with a high number of documents coded responsive.

Understanding the Cluster Visualization heat map

Applying filters or a view automatically applies heat map shading to your cluster visualization. Heat mapping helps you identify clusters that contain the highest concentration of documents matching your filter or view criteria.

Circle pack Visualization

Use the legend on the visualization panel to gauge which clusters have the greatest percentage of matching documents. In addition, you can also use the highlight matches feature. See Highlighting matches (circle pack).

Dial Visualization

Use the legend on the visualization panel to gauge which clusters have the greatest percentage of matching documents. The legend values change color when the cursor hovers over different clusters. When the cursor is not hovering over a cluster, the legend values are grayed out.

Highlighting matches (circle pack)

When setting filter conditions in cluster visualization that match a relatively small number of important documents (e.g., hot documents), the cluster heat mapping displays a very light shade of blue to indicate a low matching percentage within the cluster. Check the Highlight Matches checkbox in the heat map legend to easily identify which clusters contain matching documents no matter how small the matching percentage. Matching clusters are highlighted in orange.

(Click to expand)

Highlight matches example

Working with the document list

The document list contains documents and columns based on a combination of the following:

  • Clusters selected on the circle pack, dial and nearby clusters visualization panels
  • The view selected on the Views drop-down menu
  • Any relational items (e.g., Family) selected from the Related Items drop-down
  • Filters applied from the visualization panels
  • Field or saved search filters applied on the search panel

Sampling clusters

You can create sample data sets and then save the results as a list by clicking the Sampling icon Sampling icon . These samples take into account all the following user selections:

  • Clusters selected on the circle pack, dial or nearby clusters visualization panels
  • The view selected on the views drop-down menu
  • Field or saved search filters applied on the search panel

When sampling is applied, the cluster visualization and heat mapping is updated automatically to reflect the selected sample. For more information on how to run samples of your data, see Sampling.