There have been plethora of approaches enterprise organizations embarking journey towards Digital Transformation. The vision and strategies are revolving around leveraging technological transformations – Artificial Intelligence , Machine Learning ,Deep Learning , Natural Language processing capabilities with Proof of concepts.
Depending on the Business problem statement and the eco system (technical debt) the projects/PoC in Data Science need the amalgamation of heterogeneous technologies and thus expect to have seamless integration from an end user perspective.-
Enterprise organizations SAP S/4HANA as CORE Business systems or planned to transform to S/4HANA should now be able to leverage Machine learning building Intelligence into Business processes (sales,procurement,finance..,HR etc).
The below are the possible approaches to enable intelligence into SAP S/4HANA business processes, I will cover the insights of the approaches in upcoming blogs.
- Building/infusing intelligence into CORE -Embedded ML
- Expanding around the CORE -Consuming from SAP Cloud platform
- Consume/reuse the Services built in SAP Cloud platform
- Leverage SAP Data Intelligence to build / bring your own model / scenarios
- Expanding the CORE – Leveraging SAP – Analytics Cloud for Business user
Kindly refer below snapshot on how Enterprise Digital Core S/4 is enabling customers to leverage Digital Technologies.
Source : Open SAP Course on SAP Data Intelligence
The very first blog of mine on SAP Data Intelligence is to share the the core components of an SAP Data Intelligence we will use in building Machine Learning out of the box solutions
- Connection management,
- ML scenario manager
- Meta Data Explorer.
The below the initial screen of SAP Data Intelligence launch pad with clear focus on below points.
- Following SAP Fiori Design patterns
- Single point of entry for all SAP Data Intelligence applications
- Unified user experience
- Extensible for custom applications
Connection management: Manage seamless integration with heterogonous Data sources. This would enable us to get data insights from varied data sources – be it structured or Unstructured.
- Big Data – HDFS
- Kafka for streaming data
- AMAZON-S3 Bucket
- HANADB -HANA Data base
- DI_DATA_LAKE – Connection to DI Data lake
Please note that there are certain limitation in SAP Data Intelligence Trail account therefore we may not be able to see all the possible connections.
Meta Data explorer: Below are few options available in Meta Data Explorer
- Browse connections –
- Browse the catalogues
- View profiled datasets
- View preparations
The various options in Meta Data Explorer facilitate us to view all the connections available in the landscape, drilling down further will give us the view of the Dataset we are interested.
perform certain Data preparation activities like Missing values imputation, Data transformation, Dealing with duplicate value, Data profiling and cleansing activities
ML Scenario manager:
Organize the data science artifacts and manage the tasks centrally.
We can create Machine Learning scenarios and add artifacts
Manage model performance metrics, deployment history along with version management
A typical process within ML Scenario Manager involves:
Managing your datasets and model artifacts
Creating Jupyter notebooks for your experiments
Creating and managing data pipelines
Viewing executions and performance metrics
Tracking your model deployment
Modeler: It is based on Pipeline Engine that used a flow based programming paradigm to create data processing pipelines(graphs)
The SAP Data Intelligence Modeler provides individual Docker files to create containerized environments for the operator groups. The Docker files are selected using a tag-matching mechanism
Detailed information on various Operators, Graphs, building pipelines, monitoring the pipelines can be found in the Product help document link here
Monitoring: It provides the capability to visualize the summary of graphs executed in Modeler
The below are few of the tasks a user can perform in SAP Data Intelligence Monitoring app
- Visualize different aspects of Graph execution
- Run time analysis
- Memory usage
- CPU usage
- View execution details of individual graph instances
- View subgraph execution detail
- View Graph configurations
- Stop a graph instance execution
- Managing schedules of graphs
Reference: Product information page from SAP: SAP DI
My next blog post will explain end to end Machine Learning scenario using SAP Data Intelligence with very first use case.
For more information and support please do engage your self in SAP Data Intelligence community
Happy reading ..
Thank you very muuch