SAP Analytics Cloud provides advanced features for analyzing your data. Customers making use of SAP Analytics Cloud who use Live Connection may have questions about what happens to their data when the Time Series Forecasting feature is used. Common questions raised about the use of Live Connection in combination with the Time Series Forecasting feature are addressed below:
Q: What is the Time Series Forecasting feature?
SAP Analytics Cloud allows you to create a visualization of data plotted against a time axis, known as a Time Series chart. In addition, the user may add a Forecast to the time series chart, which makes use of SAP’s proprietary forecasting algorithms to produce a forecast into the future. The indication of the forecast quality is provided (a number from 1 to 5 where 5 is the highest quality forecast), as well as the confidence interval for the forecast as generated from the forecasting algorithm.
Q: What is Live Connection?
The Live Connection feature allows a customer to make a connection from the SAP Analytics Cloud system into a HANA system or other data source in their local (on premise) landscape, in order to retrieve data into visualizations and stories in the SAP Analytics Cloud system. Supported live data sources include BW, S/4HANA, SAP HANA, and Universes.
Q: How do I enable Time Series Forecasting on Live Connection?
By default, Time Series Forecasting will not be enabled for Live Connection. If you would like to use Time Series Forecasting on Live Connection, it is necessary for a user with the Administration privilege to enable a configuration option through the administration console. The option is called “Live Data Models: Enable Smart Grouping and predictive forecasting in Time Series Charts”.
For more details on how to enable this option please see: https://blogs.sap.com/2017/10/02/sap-analytics-cloud-live-data-connection-to-sap-bw4hana/
Q: What happens to my data when a forecast is run on a Live Connection?
When the user requests a forecast on a Time Series chart, the non-forecast data for the chart is obtained from the customer’s on-premise database in the normal way that data is retrieved to the web browser client for Live Connection. The Time Series visualization on the web browser client will first display with the non-forecasted data showing, and a progress indicator is shown. In the meantime, the Web Browser client will send a copy of the chart’s data back to the SAP Analytics Cloud back-end system. The data will be temporarily stored in a table in the SAP Analytics Cloud back-end system while the forecast is being run.
After the forecasting algorithm generates the forecasted data points, the temporary data stored in the back-end system will be removed. The forecasted data points will be sent back to the web browser client, and displayed in the chart, as shown in the example screenshot below.
Q: Does the forecasting algorithm run on aggregated, or fact-level / row level data?
The forecasting algorithm runs on the aggregated data that is shown and used in the Time Series chart itself. It is possible to change the level of granularity of the data in the Time Series Chart by clicking the icon as in the screenshot above.
Q: In which system does the Forecasting algorithm run?
The Forecasting algorithm that is used in SAP Analytics Cloud is run within the SAP Analytics Cloud back-end system itself, and does not run on-premise for a customer, or in the web browser client.
Q: Is Forecasting supported for Planning workflows as well?
Although SAP Analytics Cloud does support running Forecasting on Planning workflows in data grids, this blog post is only concerned with Forecasting on Time Series Charts. To create a forecast with planning data: view video
I hope this answers your questions about SAP Analytics Cloud regarding Time Series Forecasting and Live Connection. As a member of the SAP Analytics Cloud development team, I am excited about the value our product provides for our customers, and we look forward to hearing your success stories and feedback about how our product is useful for you. If you have any additional questions, feel free to comment below.