Using Forecasting

When an Object is generated, the system will automatically detect whether the dimension used on the first X-axis is a time-series – Which is required in order to be used for forecasting.  If the dimension can be used for forecasting, the period intervals in that time-series are determined, as shown below.

 

 

Once the system has detected that the forecasting ability is possible, two buttons, 'Step Forward' and 'Fast Forward' in the object title bar allow the user to easily forecast using a single click of a button.

Both buttons have the same basic functionality in terms of the forecasting method.  The only difference is the number of forecasted periods that each selection initiates:

§ Step Forward’ – Steps the project into the future time one period at the time, and

§ ‘Fast Forward’ – Advances forward the same number of periods as the number of existing actual periods in its data source.

Once the user clicks the Step or Fast buttons, the system automatically performs the forecasting by using the following steps:

1)  A linear regression line is added to the chart.

2)  The average percentage deviation for each actual period is calculated.

3)  The forecasted periods are added to the chart in slightly lighter colors than the actual data and with the same percentage deviation from the regression line as similar periods in the actual data.


Converting the Bar Chart to its underlying Cross Tab view, you can see where Analytics automatically inserted the additional Regression Analysis results in their own columns of the cross tab which are highlighted in red boxes, below.

If a polynomial regression is already added to the object, the polynomial line continues in the forecasted periods and each forecasted period will also have the same percentage deviation from the regression line as similar periods in the actual data, as shown below.

Note: The forecast is not limited by the time limit in the actual data, so there is no need to add extra period dimension members with no data to the data source to forecast into future or unavailable time periods.  Analytics will automatically handle the time period extensions beyond the currently available data in your forecast.

The explanation box in the object will allow the user to assess the precision of the regression as well as the coefficients of the regression line.

The traditional 'Drill Down' features may be applied along with forecasting which adds to the flexible benefits of using forecasting.