Forecasting tutorial: Previewing results

In this part of the tutorial, you preview the results of a forecasting process. The preview operation does not modify data in your Planning Analytics database.

Before you begin

This tutorial is meant to be followed in sequential order. Each part assumes that the previous part was completed.

Note: Forecasting requires that the time dimension to be placed on the column in a view. The algorithm assumes that the time dimension has no gaps. Forecasting also supports hierarchical time dimensions and nested dimensions as time (that is year, quarter, and month dimensions). Exploration views with dimensions that do not have a time dimension on the column can be forecasted, but the process might fail or the results can be unreliable.

Procedure

  1. On the Planning Analytics Workspace Home menu, click New > Book.
  2. Click Database icon to reveal the Data tree.
  3. Expand the 24retail database.
  4. Expand the Cubes node under the 24retail database.
  5. Right-click the Tutorial_<your initials> cube you created in Forecasting tutorial: Preparing data, then click Add new view.
  6. Use drag and drop to move the dimensions on the view:
    • Years on the columns
    • Members on the rows
    • Version on the context area, with the Actual member displayed

    The view should look like this:

    View configured for forecasting
  7. Select all five members on the row dimension (Member1 - Member5), then click the Forecast icon Forecast icon on the Exploration toolbar.

    The Forecast panel opens. Use this panel to set parameters for your forecast. The Scope of the forecast is automatically set to the five members you selected on the row dimension.

    Forecast panel

    As you recall from preparing the cube for this tutorial, actual values exist for all years up to 2013. The forecast should start at the first year for which values do not exist, 2014.

    Forecast period
  8. Click Select start period, then click 2014 to set the Forecast start period.

    The Forecast end period is automatically set to the last member in the dimension that is used to define the forecast period, in this case 2019.

  9. Enable the Save statistical details as comments option.
  10. Click Preview.

    The preview shows a forecast chart for the first member in the forecast; in this case it's named Member1. The chart includes forecast values for the years 2014 through 2019

    Preview of forecasted data

    You can click the Forecast selector to preview a forecast for a different member.

  11. Click Statistical details to review the details of the accuracy and statistical model for the preview.

    You can hover on any of the measures or parameters in the Accuracy section to view full definitions. For more information about the statistical model, see Statistical details explained.