Keeping corporate data up-to-date and accessing it in real-time is now a reality for SAP Analytics Cloud users. With Smart Predict and SAP HANA live data, you don’t have to deal with data uploads or updates. You can focus on augmenting your live BI stories with accurate predictions and keeping them up-to-date with a few clicks.
To benefit from this feature, you simply need the help of your administrator to fulfill some prerequisites and create connections between SAP Analytics Cloud and SAP HANA.
To go further:
Once your administrator has fulfilled the requirements and has provided you with all necessary authorizations, you can start creating your predictive model on top of your SAP HANA data in 4 steps:
To train a predictive model, you need a dataset containing historical information.
From the SAP Analytics Cloud main menu, select Create->Dataset and choose “Data from a Datasource” option. Select the SAP HANA live data repository and the SAP HANA table that contains the data you need. And that’s it, your dataset is ready!
To go further:
Now that your dataset is ready, you can create your predictive model.
If you are satisfied with your predictive model’s accuracy, you can generate your predictions by using the Apply menu.
To go further:
Once you have generated predictions on a live dataset, predictions are stored in tables in the write-back schema of your SAP HANA on premise system. To consume these predictions in SAP Analytics Cloud, you need the help of your administrator to create at least one calculation view on top of these data.
To go further:
Before you can create a BI story, you need to create a live BI model on top of this calculation view.
To go further:
As the data stays in your on-premise SAP HANA system, the data is available in real time. Any data changes you make, to your tables and SQL views in your SAP HANA on-premise system, appear immediately in the live dataset stored in SAP Analytics Cloud, as long as you don’t change the list of columns used.
To update the predictions, you need only two steps:
Access your SAP HANA live data and generate a live dataset
To benefit from this feature, you simply need the help of your administrator to fulfill some prerequisites and create connections between SAP Analytics Cloud and SAP HANA.
 | As the data stays in your on-premise SAP HANA system, an administrator needs to create a data repository that references your on-premise SAP HANA systems, and a schema that allows data to be written back to SAP HANA. This write-back schema will be used by Smart Predict to store predictions in the form of tables. |
To go further:
 | Setting up Live SAP HANA Data Access for Smart Predict |
 | SAP Analytics Cloud: Live Smart Predict with SAP HANA |
Create predictive models on top of your SAP HANA data and generate predictions you can reuse in your BI stories
Once your administrator has fulfilled the requirements and has provided you with all necessary authorizations, you can start creating your predictive model on top of your SAP HANA data in 4 steps:
1. Create a live dataset
To train a predictive model, you need a dataset containing historical information.
From the SAP Analytics Cloud main menu, select Create->Dataset and choose “Data from a Datasource” option. Select the SAP HANA live data repository and the SAP HANA table that contains the data you need. And that’s it, your dataset is ready!
| As the data stays on-premise, there is no storage in the cloud. Instead, you now have a way to point to the actual data sitting in the SAP HANA on-premise system. |
To go further:
| Creating a Live Dataset |
2. Create a predictive model and generate predictions
Now that your dataset is ready, you can create your predictive model.
- Create a predictive scenario using the main SAP Analytics Cloud menu: Create -> Predictive Scenario.
- Add a predictive model to your scenario: Select the live dataset you’ve created in step 1 and decide on the settings you will use to train your predictive model.
- Run the training.
If you are satisfied with your predictive model’s accuracy, you can generate your predictions by using the Apply menu.
To go further:
| Creating a Predictive Model |
| Look for the best Predictive Model |
| Operationalize Your Predictive Model |
3. Create calculation views to use your predictions into BI stories
Once you have generated predictions on a live dataset, predictions are stored in tables in the write-back schema of your SAP HANA on premise system. To consume these predictions in SAP Analytics Cloud, you need the help of your administrator to create at least one calculation view on top of these data.
To go further:
| Creating Calculation Views to Consume Live Output Datasets |
4. Create a BI story with your predictions
Before you can create a BI story, you need to create a live BI model on top of this calculation view.
- From the main menu, select Create -> Model. Select Get data from a datasource -> Live Data Connection
- Select the calculation view as your data source.
- Save your BI model once ready.
- And then, create a BI story using this model as data source.
To go further:
| Consuming Your Live Output Dataset in an SAP Analytics Cloud Model |
Easily keep your predictive models and BI stories up-to-date
As the data stays in your on-premise SAP HANA system, the data is available in real time. Any data changes you make, to your tables and SQL views in your SAP HANA on-premise system, appear immediately in the live dataset stored in SAP Analytics Cloud, as long as you don’t change the list of columns used.
To update the predictions, you need only two steps:
- Retrain your predictive model if you want to get it updated.
- Re-apply the updated predictive model to get the output SAP HANA table updated. The updated predictions will be immediately visible in your stories
| When re-applying the predictive model on updated live data, make sure to request the same output columns and to overwrite the previous output dataset. This will ensure that you get the data updated in all SAP Analytics Cloud features that are using this dataset! |
- SAP Managed Tags:
2 Comments
