Time Series Analysis: Process
The process consists of the following steps:
You set up the rules for the analysis in the Manage Forecast Automation Profiles app.
You schedule and run the time series analysis in the Application Jobs app.
The properties of time series data can be useful inputs for several automatic and manual forecasting processes. In the current version of SAP Integrated Business Planning, they can be leveraged automatically by XYZ segmentation only.
If you save the properties in an attribute and other results in key figures, you can use those attributes and key figures for limiting and filtering your planning views in the SAP IBP, add-in for Microsoft Excel.
You could modify the values of time series property attribute in the add-in for Microsoft Excel if you wish. However, that is not recommended because the changes you make do not influence the automated forecasting processes.
In addition to filtering planning views, you can also use the saved attributes to filter the set of available planning objects in the Assign Forecast Models app. This app allows you to assign the most suitable forecast model to each planning object.
Currently, this is the only section that you can use for setting the rules for times series analysis. The analysis can assign the following properties to any planning object in the selected planning area:
Trend - A continuous time series in which the values consistently increase or decrease.
Seasonality - A continuous time series in which the values increase or decrease repeatedly with a certain frequency.
Irregular - A continuous time series with high volatility, but no specific pattern.
Intermittent - A time series that contains intervals of various lengths between the subsequent sale dates.
Lumpy - An intermittent time series in which relatively large deviations can be observed.
Continuous - A time series with few missing values and no other patterns.
You can use more categories during the analysis. These categories define the level of details by which the system assigns properties to the analyzed planning objects.
You can choose one or more of the following options by setting the respective buttons to on:
Separate Categories per Type of Seasonality:
The system not only assigns the seasonality property to planning objects, but also the type of seasonality, which can be multiplicative or additive.
Separate Categories per Length of Seasonal Cycle:
The system also saves the length of the seasonal cycles in addition to the seasonality and, if chosen, the additive or multiplicative properties.
Separate Categories per Trend Direction:
The system not only assigns the trend property to the planning object, but also the direction of the trend, which can be upward or downward.
Separate Category for Lumpy Demand:
The system calculates a CV Squared value that shows the volatility of intermittent time series data. If the CV Squared value is higher than the threshold you specify in the next field, the lumpy property is assigned to the planning object.
The categories assigned to a planning object are combined to form property groups such as intermittent with additive seasonality (6) or multiplicative seasonality with trend.
You can choose attributes and key figures for saving the results of the time series analysis at various calculation levels. Choosing an output is not mandatory because the system saves all properties in the background anyway.
Note
If you save the results in specific attributes and key figures, you can use them for limiting and filtering your planning views in the SAP IBP, add-in for Microsoft Excel and for filtering the set of available planning objects in the Assign Forecast Models app, which can help you assign the most suitable forecast model to each planning object.
