This article will show you how you can assign relationships between measures in narratives about multi-measure visualizations. You can also check out our Relationships video here.
The three types of relationships are: ‘Portion,’ Actual vs. Benchmark' and ‘Current/Most Recent vs. Previous Period”'. You can access measure relationships by clicking on "Relationships" in the narrative settings.
Portion Relationships (continuous or discrete story types)
Portion Relationships can be applied between measures, by telling the extension that one measure is a portion of the other (e.g., “Central Region Revenue” is a component of “Total US Revenue.”
In the “Relationships tab,” click "Portion" to identify the related measures, and use the drop downs to specify that one measure is a portion of another. Once the relationship has been applied, it will look like the below.
After you've selected the parent measure, your relationship will be saved. You can click the "x" on the right side in order to remove this relationship. Or, if you had more measures, you could add more Portion relationships by using the drop downs.
Actual vs. Benchmark Relationship (continuous and discrete stories).
Another type of relationship between measures could be when one measure is meant to be the benchmark for the other. Examples of this could be comparing the total return of a mutual fund vs. some index or comparing actual sales to a sales target. Applying the ‘Actual vs. Benchmark' relationship will result in a story without unnecessary content and a focus on what's most important when comparing a metric to its associated benchmark.
Click "Actual vs. Benchmark" to identify the related measures, and use the drop downs to specify that one measure is a the benchmark for another.
After you've selected the parent measure, your relationship will be saved. You can click the "x" on the right side in order to remove this relationship.
Current/Most Recent vs. Previous Period.
In this example, we have a bar chart that is measuring the same metric, but across two different time periods. While the bar chart is across a discrete x-axis of product groups, the goal of the chart is compare the performance of those products over the two time periods. We want to answer questions such as: Which product's revenue grew the most? Which product's revenue fell the most? Using this relationship type results in a narrative containing this analysis and insight.
In the 'Relationships' tab, you will see the available relationship types you can apply. Click 'Current/Most Recent vs. Previous Period' to identify the related measures and supply other information about the relationship.
Use the drop downs to select which measure reflects the current/most recent period and which reflects some period in the past. After inputting the measures, you can input how many 'periods' in the past the previous measure is and how those periods should be referenced. The completed relationship will look like the below:
You can click the "x" on the right side in order to remove this relationship. Once you click the back arrow in the upper-left, a new narrative will be generated. This relationship can also be applied to discrete narratives.
Applying More than One Relationship
Users can now apply both 'Actual v. Benchmark' and 'Current/Previous' relationships at the same time on datasets.
Any measures that are in the visualization that are not part of the relationship configuration will be discussed in separate drill down paragraphs.
The 'Compare to Benchmark' and 'Time Period Comparison' relationships require certain scenarios in order to be activated. For example, a metric can only be compared to a benchmark if the two measures have the same 'value type', or the time period comparison can only occur in a discrete analysis narrative. The 'Relationships' window within narrative settings will provide the user with a checklist of these conditions.