The purpose of this wiki is, to illustrate with an example for planning users who would like to prepare their annual budget, how to take advantage of the existing predictive (Forecast) capabilities offered by SAP Analytics Cloud.
When starting creating his budget, commonly the planning business user starts to copy the Actual data in a Budget version and to modify existing data accordingly to the new year objectives of his company.
Let try another option and create data with the Forecast algorithm proposed by SAP Analytics Cloud.
Step by step approach
1- Create a new story, add a canvas page, select table
2- Create a new version for the Budget data (in following the step by step explanations of this wiki https://wiki.scn.sap.com/wiki/display/BOC/SAP+Analytics+Cloud+%3A+tips+and+tricks+about+Version)
3- Select time period you'd like to populate with predictive data (in our example the year 2018 hasn't any data yet) and don't forget to check "unbooked data"
4- Select the account you'd like to apply the Forecast algorithm to and click on the icon "Forecast"
5- Set up the Forecast calculation conditions
In our example, we 'd like a forecast on a montly basis starting from January 2018, based on existing data starting January 2015.
The Forecast is going to be populated until July 2018, in order to generate data for the whole year 2018, we need relaunching the Forecast a second time,
How are the number of forecasted periods generated ?
The number of periods you are able to generate depends on the number of existing periods of the dataset. The calculation rule is following:
- number of periods < 3 then forecasted periods = 0
- 3< number of periods < 8 then forecasted periods = 1
- number of periods > 8 then forecasted periods = (number of periods-8) * 25%
In order to get 12 forecasted periods in one shot, you'd need 56 existing periods (12/25% + 8) in your dataset, for 6 forecasted periods, you'd need 32 existing periods (6/25% + 8), etc....
Please note that the reference period proposed for weighing is always the previous month of the starting period : in this use case, we use December 2017.
6- Click on Preview
Please note that the quality of the forecast dataset has been noted as excellent (5/5) when the dataset was tested before creating the values.
7- Click on OK , forecasted data are generated until July 2018 and highlighted in yellow
8- To get forecasted data until the end of the year, relaunch the process starting from July 2018, please note that the algorithm proposes a new value for July 2018.
9- Now the whole year 2018 has been populated with predictive data,
You can now start analysing and adjusting your budget data to the 2018 objectives, for a smoother analysis please uncheck "unbooked data" in the Time dimension and select the 2 previous years.
Do you want to go further on the predictive topic and under the hood of the Automatic Forecast algorithm? Please refer to this blog
Use this structure to help you compose your contributions for WIKI and at the same time will ensure spelling and grammar.