Let's get in touch.
Norbert Kytka, Headquarters PlattlingContact
Clickstream analysis for an online store can take up large volumes of disk space. Many enterprises are turning to services such as Amazon S3, Google Cloud Storage or Microsoft Azure for storing these high data volumes. The data can quickly pay for itself if used correctly in SAP BW, for example, using BI processes and data from the SAP systems to understand consumer preferences.
However, actually achieving this usually involves extra steps. Essentially, the issue is data integration. A recent study by market analysts BARC found that over half – 54% – of BI professionals feel overwhelmed by the vast seas of data rolling over them. It takes considerable technical resources to bring together data from a huge range of sources and structure it clearly and harmoniously in the analysis landscape. Data from sources external to the SAP system has to be loaded manually before it can be integrated and processed. There are also numerous manual steps involved in preparing and cleaning this data before integration.
In response to these challenges, SAP SE has developed its own data management solution, SAP Data Intelligence. We’ve embraced this solution as a tool that companies can use to link together different data sources and coordinate unrelated data repositories. The integrated information is then available for further processing – thus establishing a basis for effective, decision-making with a focus on creating value.
SAP Data Intelligence succeeds the SAP Data Hub, but offers far more functionality than just integrating data from a fragmented IT landscape. For example, there are tools for searching data sources in advance.
The integrated Metadata Explorer provides a way for users to preview data in the linked sources. This means the user can examine data sources even when the data isn’t stored in an organized file system. The data exploration tools can be used to quickly identify appropriate data for a particular business process case and the example data view makes it immediately clear if there are values missing or empty entries.
Another benefit of SAP Data Intelligence is the ability to integrate machine learning (ML) into analytics processes in an automated, scalable environment. To date, it’s always been very resource-intensive to make ML-based projects productive and integrate them into IT infrastructures and operative reporting. Now, though, this is all changing with a service embedded in SAP Data Intelligence: ML Scenario Manager can be used to manage all the steps involved in an ML project.
For example, data from the Adobe Marketing Cloud, stored in Amazon S3, can now be prepared or joined in a single step. A second step trains the prepared data and then the model is made available to the end user. This has the benefit that the information can automatically be used by e.g. Fiori apps.
We’ve found SAP Data Intelligence to work extremely well in field tests. Our overall impression from what we’ve done with it so far is that it has great usability, and everything has been kept simple. Of course, the devil is in the details. If you have custom requirements for a solution, then of course you can incorporate more of your own code.
However, the idea is that users will be able to use the predefined modules to put together workflows. The interfaces and modules included make it quick and easy for users to implement data analysis projects to achieve specific objectives.
SAP Data Intelligence’s user friendliness is founded on the incorporated low-code development platform. Low-code platforms make it easy for non-IT experts to develop their own solutions. The SAP Data Intelligence platform includes over 250 predefined operators that can be used to build workflows for processing data.
Integrating ML into the workflow opens up new opportunities for many companies to make use of innovative data analysis technology. After all, it’s rare for companies to have their own dedicated data science departments, where ML is laboriously integrated into existing IT structures. Now, however, there’s an alternative way to integrate ML: by automating machine learning workflows – from data preparation to model selection, validation and provision – right out of the box.
The new models eliminate the resource-intensive work of connecting the actual data and integrating the ML model into the existing infrastructure. The data scientist no longer has to spend time working as an “integration expert” – they can concentrate on their real work of developing and improving the models.
To sum up, SAP Data Intelligence provides a basis for all kinds of exciting use cases. At the beginning of this article, we mentioned storing data from an online store in the cloud and analyzing it for the sales department. Another use case is archiving BW data from the SAP system in the cloud in the form of flat files. Or if a company wants to archive its data in a Google, Microsoft or Amazon cloud to save space in SAP BW, it can now define “pipelines” that automatically transmit data from an SAP system to, say, an Amazon data lake.
We see the new automation as offering huge potential for end-user companies, particularly due to the ability to build ML techniques directly into data integration workflows. SAP Data Intelligence also forms an excellent basis for other use cases, including production-related situations such as capturing machine data and proactively identifying maintenance requirements (“predictive maintenance”).
One thing is already clear: SAP partners will be at the forefront, working closely with SAP SE to advise their customers as they explore the new opportunities. Our experts will be delighted to have the opportunity to demonstrate the possibilities of SAP Data Intelligence and work with you to leverage the treasure troves of data in your enterprise – wherever they are.