Simulations

Overview

A simulation is a job that executes the analytics contained in the definition of an object that has been registered in the platform. For instance simulating a registered Data Element means executing the definition of that data element on a sample of records to be specified while creating the simulation job. Similarly simulating a feature means executing the analytical definition of the feature on a sample of records.

Where is this done?

Simulation can be run once an object is created by using the Run button.

Run Simulation

Refer Running Job section for details on how to run the simulation

• A simulation that has just been submitted will be marked as NEW. After clicking Run in the previous step, the Jobs page is displayed where the details on the status of the new simulation will be displayed.

• A simulation that is still being executed will be marked as RUNNING

• A simulation that has been completed will be marked as COMPLETED

• A simulation that has been failed to complete will be marked as FAILED

• A COMPLETED simulation can be stored for future reference by clicking the Save link under the Action column on the row of the respective simulation on the right side of the screen. In case the system identifies computational issues (i.e. division by zero in the aggregate definition within Data Vault) it will sometimes complete the simulation, if possible, and will provide a Logs link to the User allowing him to check the relevant errors.

Running Additional Dashboards

When running a simulation, in most cases - users can choose the dashboard they wish to see for that simulation. But in some cases, it may be useful to see dashboards about the object as well as run dashboards for the inputs to the object.

When running a simulation, users can decide if they wish to see additional dashboards by selecting the appropriate option in the DashboardView Selection -> Run Additional Dashboards section.

There are 3 options in Run Additional Dashboard(s):

• "None" (default) - No additional dashboards to be run. Just the dashboard for the object being simulation should be run.
• "Draft Inputs" - Dashboards for all the inputs which are in Draft status will be run. This is useful during Approval to quickly run simulations for all inputs that need to be reviewed.
• "All Inputs" - Dashboards for all the inputs being used (directly or indirectly) in this object should be run.

Here, "Inputs" mean all objects present in the precedent lineage of the simulation's current object. This does not include: Dependent Variable, Training Dataset, Subobjects like Transform, Strategy, etc.

Each input for which a dashboard is run will have a new Simulation Job in their page which is triggered due to their parent object's simulation.

This option is much more efficient than running a simulation for each input object one by one, because it can reuse the calculations done for the inputs while simulating the main object. So, it is recommended to use this option for faster results.

Note: Only after this simulation COMPLETEs will the simulations for the inputs be created.

Examples of simulation charts