This blog is part of a series which is related to setting up Smart Data Streaming, IoT Services & Predictive Services.
This blog is part of a series which is related to setting up few HCP services to capture event streams and predict on the data.
SAP HANA Cloud Platform, predictive services (also known as machine learning services) have been available on the HANA Cloud Platform and recently also been made available on all trial accounts. As the name suggests, it all about providing predictive analytics insights using this service. One of the main benefits of using this service is that it comes with a bunch of REST APIs which allow us to expose Predictive functionalities in your applications. As a developer, I can use any programming language to consume these REST APIs.
Using HCP Predictive services, we can now perform forecasting, find key influencers, detect outliers, and run what-if analysis.
- HANA instance with SAP APL 3.0.1 installed on this instance. The HANA MDC databases which come in the trial instances automatically have APL 3.0.1. You do not need to install anything in the HANA MDC trial instances
- The SAP HANA Cloud Platform, predictive services deployed on your HCP account. This service is also available in HCP trial instances
Installing SAP APL
Navigate to the HANA instance from the Persistence menu. If you do not have APL 3.0.1, follow the below steps.
Click on “Install Components” to add APL
Select “SAP Predictive Analytics APL” from the options and click on continue.
Setting up an APL user
Navigate to the tile “Manage Roles & Users” under SAP HANA Cockpit to create a new user called APL_USER. Since I am going to perform predictive analysis on the SENSOR table created in HANAUSER1 Schema, I am also giving privileges to this user to access the SENSOR table.
Switch to the catalog view and apply this SQL command to grant more APL privileges.
Grant the SAP APL right sap.pa.apl.base.roles::APL_EXECUTE to the newly created user.
Deploying the Predictive Services
From the services menu, locate “ Predictive Services”.
Click on “Deploy Predictive Analytics Services”
Provide the compute units and password for the account and click on “Deploy”.
This will provide a new Java application called “aac4paservices” which would be stopped by default.
Within this application, navigate to “Data Source Bindings” and add a new binding. In the below screen I have added a binding for this service to a database called sds01 with user APL_USER.
Navigate to the Overview menu and start the Application.
In the next part, I will show how to apply predictive services on the IoT table which I used earlier.