I attended an SCN SAP Predictive Analysis and SAP Visual Intelligence Webcast yesterday. First up was SAP Visual Intelligence (#SAPVisi hashtag) which I wrote about yesterday here.
First you can register yourself for all of these BI Webinars at
http://scn.sap.com/community/webinars/business-analytics
They are open to everyone.
The next open webcast is July 18th with Ingo Hilgefort on SAP BusinessObjects Business Intelligence 4.0 – Integration Update. Register here
SAP’s Dr. Adbourahmane Faye talked about “SAP Predictive Intelligence – Transforming the Future with Insight Today”. The usual disclaimer applies that planned items are subject to change.
Figure 1: Source: SAP
Dr. Faye spoke about the scenarios in Figure 1 covering data mining, the product & the solution. He covered that predictive analysis covers a range of analytic techniques.
The Gartner definition is “the process of discovering meaningful new correlations, patterns, and trends by sifting through large amounts of data stored in repositories, using pattern recognition technologies as well as statistical and mathematical techniques”.
Data Mining is defined and it helps to detect patterns and correlations with data
Figure 2: Source: SAP
Figure 2 shows that tools such as ad hoc reports and dashboarding what has happened in the past or what is happening
Predictive Analysis will show what WILL happen and why did it happen
Figure 3: Source SAP
Figure 3 talks about the need to forecast sales, relationship analysis using market basket analysis, and different services buy together.
Figure 4: Source: SAP
Figure 4 shows the different use cases for industries. Dr. Faye gave the example of banking, checking for the likelihood for customers use bank or if customer is a risky loan
Figure 5: Source: SAP
Figure 5 shows the details of the use cases. Airlines want to forecast passengers for year – model what will be number of customers
Retailers want to use market segmentation for retail chain promotion strategy
The output is a decision tree that applies to each of the segments defined
Figure 6: Source: SAP
Dr. Faye said companies want to analyze why customer s leave their company, they want to analyze the circumstances for customers moving
On the right of Figure 6 covers market basket analysis, and how to cross sell product, monitor selling habits
Figure 7: Source: SAP
Figure 7 shows the positioning – 3 pillars – empower business user – data mining in hands of end users not specifically statistician. It addresses business problems, to make the user interface simple to use and learn. It will extend BI capabilities – embedded in BI4. Dr. Faye said you can call predictive model within databases. You can use a real time analytical database with Hana.
Predictive Analysis is in ramp-up
Predictive Analysis Library (PAL) is in Hana and can call from business applications
It offers Integration with open source R in HANA
Figure 8: Source: SAP
On the right of Figure 8 is R system where you store model in HANA
The second layer is Predictive Analysis layer that interacts with Hana and interact with R
On top is the custom applications developed by SAP Predictive Intelligence Organization
Figure 9: Source: SAP
Figure 9 shows the integration with BI engine, showing the data visualization of patterns of product, accessing the relational database to HANA.
It has access to BI universes.
Figure 10: Source: SAP
In a demo as shown in Figure 10 Dr. Faye showed that when buy butter & orange juice also buy sugar too
Figure 11: Source: SAP
You can embed the model in dashboard / BI4 report
Figure 12: Source: SAP
Figure 12 shows planned innovations for Predictive Analysis. It includes enriching algorithms and integration to Business Suite system
Figure 13: Source: SAP
Figure 13 shows planned BI Innovation webinars.
Question & Answer:
Q: Can Predictive Analysis work without Hana?
A: Yes – independent from Hana
SAP Visual Intelligence:
Q: Once enriched can you write back to Hana?
A: Yes you can push data back to Hana
Great Tammy! Nice wrap up..