You need to be familiar with how to build a model in Anaplan.
This module is mandatory to use PlanIQ. The procedure is similar to setting up the modules for your source data.
To create a forecast results module:
- Insert a new module with a time dimension and a list of SKUs or items, or IDs, as rows.
Make sure you use the same dimensions as the ones in your historical data module.
- Create a new list called: forecast_action_name.
In the case of an MVLR algorithm, the new list is called: forecast_feature-names.
- Add this list as a page/context selector dimension to this forecast results module.
You need this dimension so you can capture the predictions from the different forecast actions you run in PlanIQ.
- Make sure the line item name, or code, in your forecast_action_name list is the same as the name of the forecast action you create in PlanIQ.
This helps you compare your forecast results later in the process.
- Create three line items for each quantile prediction.
- Call the line items P1, P2, and P3 and make them columns.
- Use these line items to contain the forecast results for the prediction quantiles you apply when you create a forecast action.
P1 contains the forecasts for the lower quantile, P2 for the middle quantile (0.5) and P3 for the upper quantile.
- Optional: Create line items for seasonality and trend analysis (PlanIQ provides this analysis for you).
Call the line items:
- Configure an import action.
This action imports your forecast results into this module.
Next, you can set up the module so you import different sets of predictions to compare against each other.