Running a Smart Discovery

You can run a Smart Discovery to gain insights on how underlying variables influence a selected dimension or a measure within a dataset in a story.

Context

You can launch a Smart Discovery when you start a new story or open a story containing a dataset you want to explore.

Procedure

  1. To launch a Smart Discovery:
    • If you are creating a new story: go to Start of the navigation path Main Menu Next navigation step  Create Next navigation step  StoryEnd of the navigation path and then select Run a Smart Discovery .
    • If you want to run a Smart Discovery on a data set in an existing story: go to the () Main Menu, select Start of the navigation path Browse Next navigation step  FilesEnd of the navigation path and select your story. Once the story is loaded, change the mode to Edit. From the Tools menu select Smart Discovery.
  2. Configure settings in the Smart Discovery side panel.
    1. Select the target dimension or measure on which to run the discovery from the Select Measure/Dimension list under Discovery Settings.
      Note
      When you select a dimension as your target, the Set Classification Groups dialog is displayed with all the members of the dimension listed under Baseline Group. Drag all the members you want to serve as the focus of the Smart Discovery to the Comparison Group column and select OK to continue.
    2. Specify Advanced Options settings:
      • Specify which Version you want to use for the discovery analysis. Your public Actual is the default version.
      • Under Each record refers to, provide a singular and plural identifier for each row of data, such as Employees or Sales Transactions. Record and Records are the default values.
      • Use the Included Columns section to specifically determine which measures and dimensions to include in the discovery. By default, all measures and dimensions in the data set are included in the discovery scope.
        • Select Measures to display the Set Measures to Include in Discovery dialog. Choose the measures you want to include.
        • Select Dimensions to display the Set Dimensions to Include in Discovery dialog. Choose the dimensions you want to include.
        • Select OK to submit your selections.

          The number of selected measures and dimensions is indicated as well as a listing of the selected items under Measures and Dimensions.

      • Under Page Filters, select + Add Filters to further refine the discovery scope.
        An entry will appear under Page Filter for every filter you create.
        Note
        Smart Discovery analysis can handle a finite data size; use filters to reduce the data size. Currently, you cannot run a discovery analysis on more than one million cells. By filtering the data first, you can reduce the number of cells included in your dataset. The number of cells is calculated by multiplying the number of measures by the number of records.
    3. When you are ready to launch the discovery analysis, select RUN.
      Once the process is complete, the discovery results are displayed under separate pages:
      • Overview: This page provides visualizations to summarize the results for the target dimension or measure. Even if the discovery's insight quality is considered to be poor or low, this page is created, as it is just an overview of your data.
        Note
        The Adjust Measure widget is displayed in the Overview page when the target column is a dimension. Select a measure from the widget to analyze against the target dimension.
      • Key Influencers: This page lists (ranked from highest to lowest) up to ten dimensions and measures that significantly impact the target of the discovery. A summary and insight quality is provided. For every listed influencer, there are specific visualizations displayed to show the relationship between the influencer and the target. This page will not display if the insight quality is considered to be too low.

      • Unexpected Values: This page includes a table listing existing values, and the predicted values are displayed along with the other corresponding dimensions. To add or remove columns from the displayed table, select Edit Columns. This page will not display if the insight quality is considered to be too low or there are no unexpected values.

        Two additional interactive charts (scatterplot and bar/column) are provided to help visualize the difference between the predicted and existing values listed in the table. Use either the Search or Edit Columns options to focus or filter out columns from the table. When you sort the table on a particular column, the columns in the bar/column chart will reflect this sort order. Any changes in the table will be automatically reflected in the charts.
        Note
        Unexpected Values are available only for discoveries with measures as targets.
      • Simulation: This page displays a waterfall chart containing the influencers' relative contributions. To the right, a listing of the key influencers and their corresponding values is displayed. To modify a value and simulate its impact, select the value and use either the displayed slider or radio buttons to specify a new value. The impact of the new value is reflected in the chart on the left. The value of the smart discovery target column also changes. This page will display when the insight quality is sufficient and there are variables to analyze.
        Note
        Simulations are available only for discoveries with measures as targets.

Results

After you have finished analyzing the results from your discovery, you can:
  • Save the discovery as a part of the new or existing story.
  • Share the story with other users in your organization.