xP&A HXM Workforce Planning - Create Predictive Scenario

To quote SAP’s official definition, Smart Predict describes the use of SAP machine learning algorithms to explore relationships in your dataset and build a formula – a Predictive Scenario – to predict future events or trends. Currently, three types of Predictive Scenarios are available in Smart Predict: classification, regression, and time series Predictive Scenarios.

In the context of the Human Resources - HXM Operational Workforce Planning Content Package for SAP Analytics Cloud, it is possible to make use of the time series Predictive Scenario to pre-populate future planning periods with planning values. In contrast to the standard pre-population functionality which is offered, where all values from one single historical period are copied and pasted to the periods of the plan horizon, the machine learning based algorithm is able to detect patterns in historical data and make more precise and realistic assumptions about what the future could look like. This includes mirroring seasonal cycles, upward or downward trends, fluctuations and much more.

In the following exercise, you will learn which steps are required in order to create a time series forecast scenario and how to enable Smart Predict capabilities within your planning applications.

INFORMATION

• All of the applications delivered with this content package are already prepared in a way so that you can quickly make use of the Smart Predict capabilities. This also includes the Multi Actions, which are used to start the data preparation for the machine learning algorithm and the time series forecast itself.
• What it left for you to do is to create a Predictive Scenario and embed it into the prepared Multi Actions.
• Please note that a time series forecast can only be performed when enough historical data for training the algorithm is available. 5:1 is a good ratio to estimate the horizon and get predictions with relevant confidence intervals.
• Learn more about time series forecasting in SAP Analytics Cloud in the blog post Time Series Forecasting in SAP Analytics Cloud Smart Predict and Predictive Planning in Detail.
• Please also note that the time series forecasting method in the context of this content package can only be used for the aggregated planning scenario.

In the first step, it is necessary to create a new Predictive Scenario.

1. In the SAP Analytics Cloud menu, navigate to the Predictive Scenario section.

   

2. Select the Time Series Forecast predictive model.

   

3. Store your Predictive Scenario in the folder where your content package is located or any other location of your choice.

   According to the naming convention, you can call your scenario SAP__HR_BPL_IM_PredictiveScenario for instance.

   

After saving your Predictive Scenario, a new menu opens up where you can configure your forecast models. Alternatively, you can also find the Predictive Scenario by navigating to the Predictive Scenario menu as shown in the screenshot in step 1 of this tutorial.

Now that the Predictive Scenario object has been created, you need to set it up and make some configurations. Every Plan Level of your data model requires a different set-up. In this step, you will learn how to configure the predictive models for each of the pre-defined Plan Levels. Find out more about the term Plan Level in step 1 of the tutorial xP&A HXM Workforce Planning - Add a new Plan Level.

1. Create the Predictive Scenario for the Plan Level PL1 (representing a CompanyCode, Business Unit and Division point of view).

   In order to do so please enter the term PL1 into the description window and select the SAP__HR_BPL_IM_WORKFORCE data model as your time series data source.

   

   After specifying your time series data source, the menu on the right hand side expands.

2. Set up your predictive model for Plan Level PL1.

   Please use the following settings in order to setup the forecast model for the Plan Level PL1.

   Setting	Selection
   Description	PL1
   Time Series Data Source	SAP__HR_BPL_IM_WORKFORCE
   Target Measure	FTE
   Date	Date
   Time Granularity	by month
   Number of Forecast Periods	12
   Entity	CompanyCode, BusinessUnit, Division, Location
   Filter	None
   Train Using	All observations
   Until	Last Observation
   Convert negative values to zero	True
   Influencers	None

   

3. Save your Predictive Scenario.

   

   Your predictive model for the Plan Level PL1 has now been created.

   In order to create the predictive models for the remaining Plan Levels, you can use the predictive model of the Plan Level PL1 as a template.

4. Duplicate the predictive model for Plan Level PL1.

   • In order to do so, expand the predictive models menu which is located at the bottom of the page.

   • Click on the three dots and duplicate your predictive model

   

5. Select the duplicated model in the menu at the bottom of the page and rename the description of the predictive model from Based on Predictive Model 1 to PL2

   

6. Adjust the configuration of this predictive model according to the definition of Plan Level PL2.

   • Change the entities to Company Code, Business Unit and Location only as Plan Level PL2 represents a Company Code, Business Unit (and Location in case of external workers) point of view.

   

7. Save your Predictive Scenario.

   

   Your predictive model for the Plan Level PL2 has now been created.

8. Duplicate the predictive model for Plan Level PL1 or Plan Level PL2.

   • In order to do so, expand the predictive models menu which is located at the bottom of the page.

   • Click on the three dots and duplicate your predictive model

9. Select the duplicated model in the menu at the bottom of the page and rename the description of the predictive model to PL3

10. Adjust the configuration of this predictive model according to the definition of Plan Level PL3.

    • Change the entities to Company Code, Cost Center and Location as Plan Level PL3 represents a Company Code, Cost Center (and Location in case of external workers) point of view.

11. Save your Predictive Scenario.

    

    Your predictive model for the Plan Level PL3 has now been created.

12. Duplicate the predictive model for Plan Level PL1, Plan Level PL2 or Plan Level PL3.

    • In order to do so, expand the predictive models menu which is located at the bottom of the page.

    • Click on the three dots and duplicate your predictive model

13. Select the duplicated model in the menu at the bottom of the page and rename the description of the predictive model to PL5

14. Adjust the configuration of this predictive model according to the definition of Plan Level PL5.

    • Change the entities to Company Code, Business Unit, JobFamily and Location as Plan Level PL5 represents a Company Code, Business Unit, JobFamily (and Location in case of external workers) point of view.

15. Save your Predictive Scenario.

    

    Your predictive model for the Plan Level PL5 has now been created.

Now that the Predictive Scenario and the different predictive models have been created, the objects need to be integrated into the Multi Actions which are triggered from inside the planning applications.

INFORMATION:

• In contrast to ordinary Data Actions, Multi Actions are capable of performing a chain of activities across various Versions and Data Models
• Generally, Multi Actions can be seen as a sequence of Data Actions, version management steps and predictive steps
• Learn more about Multi Actions in the SAP Help Portal

1. In the SAP Analytics Cloud Menu, navigate to the Multi Actions section.

   

2. Search for the Multi Action SAP__HR_BPL_IM_WFP_SmartPredict_PL1 and open it.

3. In the next step, add a new Predictive Scenario to the Multi Action by clicking on the binocular icon on the top lane.

   

   The new step will be added at the very bottom of the Multi Action chain.

4. Click on the new Multi Action step and apply the following settings to it:

   Setting	Selection
   Step Name	Forecast PL1 FTE
   Predictive Scenario	SAP__HR_BPL_IM_WFP_PredictiveScenarioHC
   Predictive Model	PL1
   Predictive Action	Train & Forecast
   Corresponding Planning Model	SAP__HR_BPL_IM_WORKFORCE
   Version to save Forecast	Aggregated_Plan_SmartPredictTemporary

   

5. Duplicate the newly created step by hovering over the predictive step and clicking on the thee dots.

   

6. Rename the description of the new step to Forecast PL1 FTE (external)

7. Move both of the newly created Predictive Scenario steps to their correct location inside the Multi Action chain.

   • In order to do so, use the Move up function
   • Place the Forecast PL1 FTE step in third order of the Multi Action chain and the Forecast PL1 FTE (external) step in ninth order.
   

8. Save your changes.

9. Navigate back to the Multi Actions section and open the Multi Action SAP__HR_BPL_IM_WFP_SmartPredict_PL2.

   • Repeat Sub-Step 3 to Sub-Step 8 of this Step.
   • Please note that you now have to select PL2 as your predictive model when creating the predictive steps.
   

10. Repeat Sub-Step 3 to Sub-Step 8 for the Multi Actions SAP__HR_BPL_IM_WFP_SmartPredict_PL3 and SAP__HR_BPL_IM_WFP_SmartPredict_PL5.

    • For the Multi Action SAP__HR_BPL_IM_WFP_SmartPredict_PL3 choose PL3 as your predictive model in your predictive step
    • For the Multi Action SAP__HR_BPL_IM_WFP_SmartPredict_PL5 choose PL5 as your predictive model in your predictive step

Now that everything has been set up, you need to activate the Smart Predict capabilities inside your planning applications.

Per default, Smart Predict is turned off and thus cannot be triggered when pre-populating the plan version.

1. Open the SAP__HR_BPL_IM_WFP_OVERVIEW_PAGE application in edit mode.

2. On the left-hand side of your development environment, open the script getCustomizing

   

3. Set the parameter toggleIsPredictiveEnabled to true as shown in the screenshot and save your changes.

   

Now that you have set up the Predictive Scenario and activated the Smart Predict capabilities, you can run the predictive algorithm for pre-populating your plan version.

1. Open the SAP__HR_BPL_IM_WFP_OVERVIEW_PAGE application.

   INFORMATION

   It is important to open the SAP__HR_BPL_IM_WFP_OVERVIEW_PAGE application first instead of directly jumping to the SAP__HR_BPL_IM_WFP_CENTRAL_ASSUMPTIONS application as the parameter you changed in the previous step is transmitted via the overview application. If you opened the central assumptions applications directly, smart predict capabilities would not be activated.

2. Enter the Application Configuration application.

   

3. Click on the Confirm button and activate the toggle Initialize Plan Data After Publish. A new toggle Use Predictive will appear below. Activate this toggle and hit the OK button in order to run your Predictive Scenario.

   

   INFORMATION

   Both the traditional pre-population and the Smart Predict based pre-population take place now. The traditional pre-population, where values from one reference period are copied over to the periods of the plan horizon, are written on the Aggregated_Plan version as usual. The Smart Predict values on the other hand are written on the Smart_Predict version. This way, you will be able to use the values from the Smart_Predict version as a reference for your planning assumptions made on the Aggregated_Plan version.

4. After the prediction has run through successfully, click on the Home icon on the top left corner to get back to the overview application.

   

5. Open either the Aggregated External HC Plan or the Aggregated Internal HC Plan application. In this example, the Aggregated Internal HC Plan application is opened.

6. Activate the toggle Show Smart Predict Reference.

   

   You are now able to see the results of the time series forecast under the Smart_Predict plan version.

Congratulations! You have now successfully created a predictive scenario.

Interested in more xP&A topics and related business content packages? Visit our community page Extended Planning & Analysis Business Content.