Creating a what-if measure

Last modified by Aurelie Bertrand on 2026/07/01 12:10

You can create what-if measures based on measures and calculated measures. A what-if measure allows you to study or observe the influence of one or more measures on a given measure.

To create a what-if measure:

  1. Click the New Measure button, then, in the context menu that appears, select Create What-if measure.
    New measure 

    ➡ The What-if measure dialog box appears.
    Whatif_analyzis_dialog_en.png
  2. Enter the Name you wish to give to your simulated measurement.
  3. Select the Category if needed.
  4. In the ‘Measure to analyse’ section, select:
    • The Name of the measure you wish to analyse
    • The Aggregation of this measure
  5. In the ‘Dependent measures’ section, select the measures you wish to use to model the measure to be analysed.
  6. For each selected measure, enter:
    • The Aggregation of the measure
    • The Associated variable with this measure. If you do not wish to vary the  measure, select ‘None’.

Two buttons are also available:

  • The ‘Automatic Model’ button allows you to automatically select the measures needed to model the measure to be analysed. To do this, you must also select the dimension(s) according to which you wish to explore the measures in order to determine the model.
  • The ‘Quality of the model’ button Bouon_calcul_qualite_modeleallows you to calculate the adjusted coefficient of determination (adjusted R²). This calculates the model’s accuracy whilst taking its complexity into account. To do this, you must select the measures in your model, along with their aggregation, and also the dimension(s) according to which you wish to explore the measures in order to calculate the accuracy. It also indicates the measures that contribute most to the model.

💡 See the page Predictive tutorial for more details and an example.