Share

Running EPM Auto Predictions Using Qubix Cloudbridge

Predictive Planning has long been part of the EPM Cloud, but it is limited to predicting trends for the data slices on individual forms. Now, Oracle have added Auto Predict, which uses the same forecasting model as Predictive Planning, but is not limited to individual forms. Using Auto Predict you can run a prediction based on any slice of actuals data and store the results in another slice. You can then present the results using forms or dashboards. Auto Predictions tasks are saved as jobs in EPM, which means that they can be run automatically by Cloudbridge as part of a multi step process. This means that you can ensure that your predictions are always recalculated as soon as any new actuals data is loaded into the EPM model.

Creating an Auto Predict Job

To create an Auto Predict job, select Application, then choose Overview. When the overview has opened, choose the Actions menu and select Auto Predict.

Creating an Auto Predict Job

Click Create to set up a new prediction. You can give it a name and description.

Revenue prediction

Select the cube containing the actuals data that is the source for this prediction. By default, this cube will also store the prediction results. If you want to put the results in a different cube, choose “Cube to Cube” and select the destination cube.

Define the slice to predict in the “Analyze” section and the slice in which to store the results in the “Predict” section.

Select members

In the “Analyze” section you can choose how much historic data to use for the prediction. The default is 3 years. 

In the “Predict” section you can choose how far in the future to forecast. The default is six periods.

Click Save to complete the job.

Automating the Auto Predict Job from Cloudbridge

Once the Auto Predict job has been defined, you can run it from Cloudbridge using the Process Builder. In this example, we assume that you have already created an environment and system for your EPM application in the Process Builder. If you are new to the Process Builder then have a look at our free Cloudbridge training course on Qubix Academy.

From the Cloudbridge menu, choose Process Builder, then choose Configuration to open the processes diagram.

Process building

Right-click on the system for your EPM Cloud server and choose Add Process.

Add process

In the Properties Panel, enter a name for this process, then choose “qbxPBCS.runJob” from the list of commands.

Process panel

Complete the parameters for the command as follows:

jobSpec: The instructions to EPM Cloud about the job to run. Here is an example a jobSpec to run a job called “Prediction1”

{

    "jobType": "Auto Predict",

    "jobName": "Prediction1",

    "parameters": {

        "forceRun": true,

        "paginatedDim": "Entity"

    }

}

jobType is always “Auto Predict”. jobName must exactly match the name you gave to the Auto Predict job in EPM Cloud.

Leave the forceRun parameter as true to ensure that the prediction is re-run even if the actuals data hasn’t changed.

To make the prediction faster, choose a dimension in which the data is evenly spread and enter it as paginatedDim. The prediction will run in parallel across separate threads based on the spread of data in this dimension.

maxRetries: Number of times to retry connecting to EPM if there is a communication problem.

pbcsApplication: The name of your EPM application

pbcsPassword: The password to login to EPM

pbcsRestAPI: The URL of the EPM API interface. The format of this URL is https://EPM_Cloud_URL/rest/v3, e.g. https://planning-test-company.pbcs.em2.oraclecloud.com/HyperionPlanning/rest/v3

pbcsUsername: The username to login to EPM

TimeLoop: The number of seconds between each time Cloudbridge checks whether the Auto Predict job has finished

Timeout: The number of seconds to wait for a response from EPM Cloud

Once the parameters have been entered, you can test the job by right-clicking the process and choosing “Run from here”

Add new process

Where this command will be most useful is as part of a process that loads new data into the EPM application. In that scenario, you will want to place this command in the process after the new data has been loaded, to ensure that the predictions in the model have been recalculated to reflect the new actuals data.

Become a member

Newsletter

Receive our newsletter to stay on top of the latest posts.