Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
Showing results for 
Search instead for 
Did you mean: 
Welcome to the seventh installment in the PAi and S/4HANA Blog Series. In this blog, we are going to explain how the powerful combination of S/4HANA, PAi, and SAP Analytics Cloud drives better data science productivity for Intelligent Enterprises.

SAP Analytics Cloud offers a portfolio of Augmented Analytics features that help users achieve advanced insights through machine learning technology. Smart Predict, which is part of this portfolio, generates insights about what is likely to happen in the future. This information can then be shared with the business so that decision makers can confidently take the right action.

Putting Business Experts in the Driver’s Seat

Data science carries a stigma – it is seen as a highly complex, expensive, and requires a PhD to understand. Data Scientists can spend weeks or months on a predictive project and fail to deliver something that the business will actually put into production. Why is this?

Data scientists are experts in statistics and algorithms, not necessarily your business. To build the right analytical dataset, they require help from subject matter experts to understand the problems the business wants to solve, the right data to include, where the data is stored, and how the data is structured. It is not a simple task and requires repeat experimentation to adjust to the nuances of your organization. However, it isn’t always easy to achieve this type of collaboration, and data scientists often end up working in silos.

Even if data scientists are able to prove a significant pattern and generate predictive insights, we often see that the business does not trust the output. Without knowledge of what data was used, how it was manipulated, and what happens in the data science process, decision makers are weary to act upon these insights. Sadly, all that hard work and effort then goes to waste.

What if business experts and analysts could drive predictive projects themselves? With an excellent understanding of business data, and strong relationships with business decision makers, they would be in an excellent position to create recommendations for business process optimizations. The need for a simper, reliable, and trustworthy solution is exactly why SAP Analytics Cloud’s Smart Predict was created. Smart Predict makes data science possible for analytical business users through predictive automation. Our proprietary algorithms handle many of the complex steps in the data science process, ensuring business experts and analysts can move from data to meaningful predictive insights in minutes.

Focus on Business Questions, Not Algorithms

Smart Predict accelerates the prediction and recommendation creation process by focusing on business outcomes, not data science know-how. Instead of putting the onus on the user to statistically balance the analytical dataset, remove outliers, encode the data, pick the right algorithm, and select the right model, Smart Predict users can train a predictive model in as little as 3 steps:

  1. Select the type of business problem you are trying to solve

  2. Choose your analytical dataset

  3. Select your target (the column representing the historical answer to your business question

Smart Predict addresses most business problems through 3 predictive scenarios: classification, regression, and time-series.

Classification Scenario: If you’re trying to determine the likelihood of whether something will happen, you’re dealing with a classification scenario.

Regression Scenario: If you’re trying to predict a numerical value and explore the key drivers behind it, you’re dealing with a regression scenario.

Time Series Scenario: If you’re trying to forecast a future numerical value based on fluctuations over time, seasons, and other internal and external variables, you’re dealing with a time series scenario.

Once the model has been created, Smart Predict provides transparency around the model’s accuracy, robustness, key influencers, and other details a data scientist would typically consider when evaluating a model. These are highlighted in a way that business experts and analysts can understand to ensure they are comfortable with applying the model on new data.

Smart Predict helps non-data scientists create meaningful insights to guide decision making in a truly self-service manner. Now, more employees within the organization can leverage machine learning technology to rapidly move from business question to trustworthy predictions, injecting more advanced insights into your business.

Intelligence Where You Need It

Data science projects are valuable because they provide new insight that can be used for decision making to maximize our chance of future success. These advanced insights need to be accessed and understood by decision makers at the right time to ensure they influence the right action.

Many organizations utilize reports or dashboards to communicate these advanced insights to business users. While some users may be savvy enough to search for this information, many users will make decisions based on the information available within the day-to-day applications they are familiar with. We need to bring predictive insights closer to the processes themselves to ensure they are seen and acted upon.

Integration is a major focal point here at SAP, so we’ve made it simple to embed predictive insight into S/4HANA. Business experts can work with SAP Developers to create Fiori tiles which open SAP Analytics Cloud stories and visualizations. Or, they can take it a step further and publish the predictive models created by SAP Analytics Cloud to S/4HANA through the Predictive Analytics Integrator (PAi). The models can then be trained and applied in near real-time, and the predictive output can be blended with transactional data and highlighted in custom Fiori apps.

Bringing it All Together

As more organizations adopt data science and machine learning practices, it is important to remember the fundamentals. If users do not understand or trust the models, they will not be used. If decision makers are not exposed to predictive insights in the right way, you will see minimal impact on your business. Adoption of machine-generated insights requires substantial cultural changes.

By exposing more users across the organization to data science processes and outcomes, SAP is leading the way in demystifying data science to drive further adoption and productivity. We’re on the Intelligent Enterprise journey ourselves, and we can’t wait to hear about your progress!