Blending Models with SAP Analytics Cloud

[nvytla]

What is Blending?

Blending models lets you join a primary data source (such as ‘travel expenses’) with secondary data sources (such as ‘employee headcount’) that contain common linked dimensions (such as 'employee'). A new model is not created when you blend models, and the original models are not modified. Links between models that are blended only occur within a story. Blending can be done within individual tables and charts.

Some common uses for blending

• Comparing data of actuals from a corporate SAP source with plans stored from a non-SAP source.


• Comparing sales results against market benchmarks, or marketing campaign results available in .csv files, or data provided by a 3rd party consulting firm.


• Creating custom calculations based on key performance indicators from different data models like public census data.


• Using Smart Predict to output predictions in a dataset, and blend these predictions with other data sources, in the context of a story.

Technologies used for blending

There are 3 technologies used for blending:

• Native blending:

This means the blending occurs within SAP Analytics Cloud 'natively'. No additional technology is needed.

• Browser based blending:

This means the data is transmitted, via the users' web browser, and temporarily stored in either the Primary or Secondary data source where the blending itself occurs. Although no special setup is required it is highlighted as additional requirements or restrictions apply. Please refer SAP Note 2715871.

Typically recommend browser-based data blending to business users or data analysts who want to blend smaller aggregated datasets together within a story. A system-defined data volume limit ensures that the data blend operation within the story performs as expected. In general, we recommend trying browser-based data blending with live models first, because this approach should meet most business users' needs.

When users are linking dimensions for the purpose of browser-based blending, a warning is shown if the data volume limit is exceeded:

An error message is shown if the blending operation doesn't complete successfully due to the data volume limit being exceeded:

Note:

SDI and browser-based blending are defined on a per-connection basis, with SDI-based blending having the higher priority. For example, if you have both SDI- and browser-based blending configured against a connection for a data source, then SDI will be used for blending the data.

When using browser-based data blending with a live SAP HANA source, data from the live HANA system is temporarily cached in SAP Analytics Cloud.

• SDI based blending:

This means the data is transmitted, via Smart Data Integration (SDI), and temporarily stored in either the Primary or Secondary data source where the blending itself occurs. SDI need to be setup between the data source and SAP Analytics Cloud. This means an SDI Agent to be running and connected to the HANA system via SAP Cloud Connector.

SDI is recommended over browser based blending.

Matrix to show which models can be blended to other models based upon the connection underlying the SAC model

Source for above matrix

Note:

• Data blending with imported data stored in SAP Analytics Cloud don't require any blending-specific prerequisites or setup.


• Data blending with models based on live data is also supported for specific SAP data sources, but does require additional prerequisites and setup.

Additional Information

• Creating Linked Dimensions. Link


• Enabling Browser Based Data Blending for Live Models. Link


• SDI-Based Blending Between SAP HANA Live Data Models and Import Data Model. Link


• SDI-Based Blending Between SAP BW Live Data Models and Import Data Models. Link