OpenSAP Big1 Week 2 Transcript

8/19/2019 OpenSAP Big1 Week 2 Transcript

1/20

openSAPDriving Business Results with Big Data

Week 2 Unit 1

00:00:12 Welcome back to the Driving Business Results with Big Data openSAP course. This is oursecond week, and this week we're going to have five units

00:00:20 looking at Big Data from our perspective as we see it as SAP. So the first week we dedicatedto real-life examples, use cases.

00:00:29 We looked at situations where organizations, humans, individuals, or businesses solved theirBig Data problems.

00:00:37 Now, this week we're going to look at the actual technology, the solutions that we haveavailable to solve those problems.

00:00:45 Now the interesting thing about Big Data is that Big Data itself is the result of all the evolutionthat has happened in the past 20, 30, 40 years in data processing technologies.

00:00:59 And, as a result, it actually created a couple of problems. These problems are not really badproblems but we have to solve them before those problems turn into solutions

00:01:13 So, all the technology that enabled us to collect all this big amount of data pretty much fromeverything in our lives, from smartphones, from all different kinds of devices,

00:01:24 from signals, signal data from sensors, from machines - pretty much everything is beingcollected today by individuals or by organizations.

00:01:35 Now, why should we collect it if we don't really have any benefit out of it? It's just like our ownlibrary at home, for example, we have been buying books for the last decades,

00:01:46 which is kind of, you know, still manageable, maybe we manage to read all those books. Imean, why did we buy those books if not with the purpose to read and actually learn fromthose books?

00:01:56 Now, today we are capable of downloading unlimited amounts of electronic knowledge fromonline retail stores. We can buy books on our Kindle with no limit pretty much, but:

00:02:11 does it make sense to collect and download all that data, to pay all that money to get all thatdata? All that information if we cannot really turn that information into real results?

00:02:22 For that reason, we need solutions. Now this applies also to business. On a business level,companies today are capable of collecting enormous amounts of data.

00:02:33 Again, whether it's machine, sensor, social media, e-mails, text data, whatever it is, it'sgrowing every year, it's getting more and more.

00:02:44 So this week we're looking at those tools, those technologies and those SAP Big Datasolutions which actually help either individuals or businesses or communities to extract valuefrom Big Data.

00:02:58 Today's unit is kind of an overview unit. Today, we're going to cover the big picture, we talkabout how we at SAP see this whole topic of Big Data.

00:03:07 And what kind of solutions we have available for this. And then in the next four units of thisweek, we are going to go into the details of each component of Big Data.

00:03:18 Unit 2 tomorrow, we're going be talking about SAP's Big Data platform. In the next unit, we'regoing to talk about the techniques and tools available to acquire the data, collect and gatherthe data.

00:03:31 And Unit 4 we're going to dedicate to applications and analytics. We're going to be looking atthose components which help us to act, and visualize and analyze the data.

00:03:42 And, in the last unit of this week, we're going to talk about data science. We're going to look atwhat it really takes once we gather, collect, and analyze the data,


2/20

2

00:03:52 what it really takes to get the meaning, the real value out of the Big Data. So, for us, Big Datareally refers to the problems of

00:04:05 acquiring, capturing, storing, processing, and also then analyzing massive amounts of BigData. Now, typically that data is coming from multiple sources, all kind of sources, and it'scoming in all kinds of formats

00:04:19 structured or unstructured. So today, businesses and organizations are getting that data eitherfrom enterprise systems,

00:04:28 from social media, from sensors, from signals, from machines and sensors, videos, images,weblogs, e-mails, whatever it is, it all generates Big Data.

00:04:41 Now, why should we then collect and look at that Big Data? There can be only one realreason, we want real results.

00:04:49 We want to find the meaning behind the big numbers, we want to find correlations betweendifferent patterns, trends, and so on. So, businesses today, they find those meanings andthose values, real values

00:05:03 in various use cases. It could be demand, a customer demand. It could be sentimentintelligence, fraud detection,

00:05:13 predictive maintenance, or supply chain optimization. So there are all kinds of areas where BigData can help us to find real results.

00:05:23 Obviously, on the way from getting from the collection of the data to the real results there are alot of challenges that businesses face today.

00:05:32 For example, there might be old database systems still in place where it's very difficult to getthe data quickly into. Then there might be very complex IT environments, multiple servers,multiple databases, that businesses have to deal with.

00:05:47 Also, the data is very different by nature. There's a big difference between a free unstructuredtext data in the form of e-mails,

00:05:55 or social media blogs, for example, versus structured relational database data stored in files onhardware.

00:06:10 Then, if we manage to gather all this data, and we have this available, then we can have aproblem that there are no skills really available, or no tools for the experts

00:06:21 to analyze that data and get the meaning out of the data. So the lack of expertise might be aproblem. Or, once we have all the data, and we analyze the data, we capture the data,

00:06:32 we looked at the data. Then the next question might be how to distribute this data, how to givethis knowledge, this information. How to visualize it for the end user, how to give them accessto it, give them tools to get all this data.

00:06:45 And, once we even pass that hurdle, then we might have the problem that once we knoweverything, how to apply it to the business operations.

00:06:55 How to change business processes, how to change the daily operations of an enterprise. Thatmight be another challenge.

00:07:01 Now, here at SAP, we have kind of the answer, but we kind of summarize around the 3 As. So

we have tools and solutions for acquiring, analyzing, and acting upon that Big Data.00:07:13 Now this is a continuous process, it's a cycle that's going round and round. It really makes no

sense to just collect and analyze the data just once.

00:07:21 Businesses, they want this on a daily basis, on a weekly basis, data is coming in from sensorsnon-stop. We cannot just stop that data stream and they'll start analyzing, and then whenwe're ready, we'll continue with the next set.

00:07:35 This is a continuous process. Now, here at SAP, we created tools and technologies to dealwith all the phases, all the stages of this process.

00:07:44 For the acquisition part, we have data connectors to connect all kind of sources of data to getthat data.


3/20

3

00:07:51 Then, we have sophisticated ETL tools to extract, transform, and load that data into ourdatabase. Then, we also have streaming analytics to get the data streams from machines andsensors.

00:08:08 For the analyze phase, we have, again, a whole line of tools. We have advanced analytics, wehave business intelligence, and agile visualization tools.

00:08:19 We're going to cover all these in more detail in the further units. Then we also have Big Datadevelopment tools.

00:08:26 So, basically, we have a platform. On that platform, we can create models, we can createanalytical applications to then analyze and look at that data.

00:08:38 And, then, we also have applications, both line of business and also industry applications toact upon the Big Data.

00:08:48 So, with that, we have a full circle of offerings to cover this whole cycle of continuous dataacquisition, analyzation, and acting upon. If we look at the main components of SAP Big Datasolutions, then there are three major components.

00:09:06 So, first we have the SAP HANA platform for Big Data. And, then we also have the Big Dataanalytics and applications.

00:09:15 And, we also have the Big Data Sciences and the related services. So, we're going to look intomore details of those components in the coming units.

00:09:24 The first big component is obviously our SAP HANA platform, which is obviously more than just a database.

00:09:31 It is a platform both for development and deployment of those SAP applications and solutions,also including SAP Big Data solutions and applications.

00:09:44 We're going to dedicate a whole unit, actually two units, to this topic. In the next unit, we'regoing to be talking about the platform itself, SAP HANA, plus other platforms which worktogether with SAP HANA like one ecosystem.

00:09:58 And, in the third unit, we're going to cover all the acquisition techniques, how we can get thatdata into our platform. So those other components actually related to the platform, such asSAP HANA obviously, and then we have SAP IQ.

00:10:16 And, also, Hadoop, is part of this big picture as a major player in the Big Data platformbusiness globally. So, we're going to dedicate another unit to this topic when we're going totalk in more detail about the details of this.

00:10:32 As for the SAP Big Data solutions and applications and analytics, we have basicallyapplications and solutions available for all those use cases where businesses can find actuallyresults, real results

00:10:45 in Big Data. So, those things are applications related to customer demand. So, we haveapplications to analyze customer demand, to look at customer information,

00:11:00 to do audience discovery, account intelligence, for example. Then, segmentation of customers,profiling of customers, and so.

00:11:08 Then, we have applications around fraud management. There are other applications forDemand Signal Management, which is very important for the retail industry.

00:11:21 We also have Sentiment Intelligence. There are some more like Stream Intelligence, SignalIntelligence, for example. And, then we also have Predictive Maintenance.

00:11:32 And, we have everything that's related to the Internet of Things, connected manufacturing orconnected logistics. We're going to dedicate a whole unit, Unit 4, of this week, when we talkabout more details about these solutions.

00:11:45 And, on the analytics side, we have various tools. Now, some tools are designed for endusers, some tools are designed for data scientists, for experts.

00:11:56 For example, we have SAP BusinessObjects business intelligence tools to kind of engageorganizations and the employees of the organization in the analytical process.

00:12:11 Then, we have SAP Lumira, for example, on the visualization front where we can create verysophisticated charts and graphs, including geospatial information, for example.


4/20

4

00:12:23 And, for the real experts, the data scientists, we have predictive tools, such as SAP Predictive Analytics and SAP InfiniteInsight. So, we're going to dedicate a whole week to the topic ofpredictive analytics.

00:12:36 That's going to be our Week 4 of this course, when we're going really look at the details of thissolution. And, then, we're also going to have hands-on experience. How to create predictivemodels in these tools.

00:12:51 Then, the last component of the entire Big Data solutions and offering by SAP is the datascience and data science services that SAP has. Again, we're going to cover used cases.We're going to talk about examples.

00:13:05 And we're going to really look in Unit 5 of this week at what makes data sciences different fromtraditional analytics. So, stay tuned, we're going to get to that point.

00:13:18 And, to close this unit, we also have to emphasize that there are SAP Rapid Deploymentsolutions. Now, the SAP Rapid Deployment solutions are those pre-packaged solutions

00:13:31 which help our customers to adopt all these technologies. They include standard software,very extensive documentation,

00:13:41 user guides, manuals, and also very often pre- configured content, which then helpscustomers to - both in terms of time and cost - adopt those solutions much faster.

00:13:57 So, we're going to dedicate three weeks to this topic. We're going to talk about SAP Big DataIntelligence Rapid Deployment solution in our Week 3, that's going to be our next week.

00:14:06 And, then, we're also going to have SAP Predictive Analytics, Week 4. And, we're going tofinish this course in Week 5 with the SAP hybris Marketing Rapid Deployment solution.

00:14:16 The structure of those weeks is going to be very similar, we're going to be talking about firstthe product, the related product, so we're going to have guess speakers for both predictiveanalytics and hybris marketing,

00:14:27 where the product management will tell us all about these products. And, then, we're going tolook at the overview, what the rapid-deployment solution really has inside,

00:14:38 how to deploy it, and we're also going to have hand-on system exercises as we talked about inthe previous week. And I'd just like to remind you this time that

00:14:50 please make sure that you're getting those users and accounts for AWS Amazon WebServices and also SAP CAL Cloud Appliance Library because we are one week away from thefirst Big Data Intelligence week

00:15:10 when we're going to actually do hands-on exercises. So, with that, I'd like to summarize whatwe've covered today.

00:15:19 We talked about our SAP Big Data solutions. This was kind of an overview, so it's not all.

00:15:26 We're going to now spend four more units to go into details on each of these components.Stay tuned and I'll see you again in Unit 2.


5/20

5

Week 2 Unit 2

00:00:12 Welcome back to the Driving Business Results with Big Data openSAP course. This is Unit 2of the second week, and as you will remember in the first unit we kind of did an overview

00:00:24 about all the SAP Big Data solutions that we have available. We looked at the components ofthose. And, in this unit today, we're going to start drilling into the details of the first majorcomponent,

00:00:35 which is the SAP Big Data platform both for storage and computing of data. So, let's just for amoment take a step back and let's look again at those major components.

00:00:50 So, we have, on the one hand, the SAP HANA platform, which is the topic of this unit and alsothe next unit. This is what we use to acquire and also store and then process the data.

00:01:02 So, then we have the Big Data apps and analytics. That's to analyze and act upon the BigData.

00:01:09 Then we have the line of business and industry applications and also the analytics to gaininsight and visualize the data. That's going to be our Unit 4 of this week.

And then we have data services and data science services,

00:01:26 which is the services of our experts, our data scientists, to actually find the real meaning, thereal value behind Big Data. So that's going to be our Unit 5 of this week.

00:01:39 So first let's talk about out the SAP HANA platform. As we've mentioned before, SAP HANA isreally more than just a database.

00:01:48 It is SAP's flagship in-memory processing product. It can do much more than just store data. Itis a platform.

00:01:56 It is a platform for both developing and also deploying high-performing, high-value analyticalapplications, solutions, including SAP Big Data solutions.

00:02:08 So the great thing about SAP HANA really is that it converges database, data processing, andalso application platform into one.

00:02:18 So, in the past, we know that there has always typically been a separation betweentransactional platforms and analytical platforms. That was because we didn't have the suitabletechnology

00:02:30 to actually provide one single platform for both analytical OLTP and also OLAP processes,analytical and transactional processes. But in HANA we see that in one server we can have aplatform for transactional processes,

00:02:50 for analytical processes. We also have tools built in SAP HANA for predictive analytics. Thereis a predictive analytical library in SAP HANA which lets us do predictive analysis,

00:03:05 to build predictive models within HANA without using other tools. There are other, by the way,alternative tools as we are going to see in Week 4 of this course.

00:03:14 But SAP HANA by itself is capable of doing predictive analytics. Then we have planning tools,and we also have text processing tools and spatial tools.

00:03:23 Now, about the text processing, which is an interesting feature here, SAP HANA's textprocessing capabilities is built-in, HANA has it built in itself.

00:03:34 And with that, we are capable of gathering all the data from social media or any other kind ofunstructured text-based sources, and SAP HANA can actually categorize and group thosedifferent social media posts

00:03:52 according to sentiment, for example, of those posts. So all these capabilities are built in. Thenwe also have Hadoop.

00:04:01 Now, why are we talking about Hadoop here? I mean, obviously, Hadoop is kind of thebenchmark, the biggest, the first probably Big Data platform available on the market.

00:04:14 Now, it's very different from SAP HANA and that's why it's really important to put it side by sideand talk about what are the advantages of Hadoop and why we actually rely on this platformas well


6/20

6

00:04:27 in our SAP Big Data solutions. So, first of all, it's an open source project. It's built to manage areally huge amount of data.

00:04:36 We're talking about pretty much infinite storage capabilities, or multi-terabyte or petabyteamount of data. Hadoop comes together through a network of commodity servers.

00:04:51 So because of all these features, Hadoop is capable of scaling and performing at a relativelylow cost. This is one of the big advantages of Hadoop.

00:05:02 It complements really well the traditional data warehousing and data analytics processing but itdoes not replace it.

00:05:10 In our opinion, it's a very good match, to put, for example, SAP HANA and Hadoop togetherand use those different platforms for different processes. We're going to talk about this in moredetail soon.

00:05:23 So, Hadoop is really good at many things. It's good at massive scalability as we said earlier.

00:05:31 It's good at storing file-based data, for example log files, Web logs, and so on.

00:05:39 Then also it's a very good recipient of sensor data because of the large volume of that data.It's also very good at processing, storing, and processing unstructured data

00:05:52 or multi-structured data. So we can dump anything in Hadoop which is not in structures andalso typically data that has no schema.

00:06:02 So that's one of the differentiators from SAP HANA where we store data in schemes, inschemas.

00:06:13 And in Hadoop, we really don't need that capability. Then we have SAP IQ.

00:06:20 Now SAP IQ is actually a very interesting database and platform as well. It's in fact the firstcolumnar SQL database available. It's been around for 20 years like that.

00:06:33 And, because of its columnar nature, it has extraordinary compressing technology built in. So,by the way, SAP IQ holds two records in the Guinness Book of Records.

00:06:48 On the one hand, it is the fastest data load ever recorded. It was 33.4 terabytes per hour,which makes about one petabyte per day of data loaded into the system, into SAP IQ.

00:07:05 So that basically tells us how exceptional the bulk data load capability is of this platform. Andthen also the most data ever stored in any database.

00:07:20 SAP IQ holds the world record. It is 12.1 terabytes of data stored. And now here we come tothat feature of SAP IQ about the compression technique that is used here.

00:07:35 So, in SAP IQ we can actually do 75% compression. Another very useful feature and why webelieve that it has to be part of this whole ecosystem

00:07:45 is because SAP IQ integrates really well with Hadoop. It can receive and read data very well,communicate with Hadoop very well.

00:07:56 So now let's look at these three platforms. Three separate database systems or platforms,which, if we put them together, we can use them according to their strength.

00:08:07 And we can select the right platform according to the source data. So, they have differentstrengths.

00:08:14 For example, on the left side, on the side of SAP HANA, we're talking about real time and

really high speed, so that's the advantage. We are talking about an in-memory platform afterall.

00:08:26 On the right side, we have the infinite storage, Hadoop. Unlimited, pretty much unlimited,numbers scalable, low-cost storing capabilities. In the middle, we have SAP IQ, which is kindof in between.

00:08:38 As we've referred to it, it's not hot, it's not cold, it's warm. So we can characterize data like that.

00:08:46 Typically SAP IQ we can use for situations where we don't really want to put data on cold. Andwe still need to have access relatively frequently to the data, and it's a very large amount ofdata.

00:09:01 So SAP HANA, which could do the job, but we don't probably have the resources and we don't

want to spend so much cost on deploying a larger HANA system.


7/20

7

00:09:13 So, we can side by side deploy an SAP IQ system, which because of its extraordinarycompression technique can actually store a very large amount of data.

00:09:23 So we sometimes refer to SAP IQ as a near-line storage. So, obviously, SAP HANA has anadvantage that

00:09:32 it has a capability of creating virtual tables and it has a virtualized access to any data source infact. So we can actually use integration between Hadoop and HANA or SAP IQ and HANAwhen data is still in those remote sources.

00:09:53 But through virtual tables, we can get access and we can process the data. And Hadoop, onthe other side, is typically designed for batch-based, batch processes,

00:10:06 offline data storage, and also for archiving reasons. Now the next topic we're going to talkabout, actually we dedicate a whole other unit to this.

00:10:21 But let's just summarize what it is. It's the way that we acquire the data. So until now we'retalked mostly about the storage of the data and then processing that data.

00:10:31 Let's say a few words about the data acquisition components here. So, obviously, here wehave a lot of challenges.

00:10:39 The challenges are that Big Data is coming from an infinite amount of sources. We typicallyhave in our SAP Big Data solution ecosystems, we typically have SAP systems

00:10:53 like SAP Business Suite components: ERP, CRM, SRM, SCM, for example. Then we alsohave our data warehouse, SAP Business Warehouse, where we want to retrieve data from.

00:11:06 And we can also use SAP IQ as another database to get the data from. Now for thesecomponents we typically use either real-time, trigger-based replication techniques,

00:11:18 such as SAP LT, or SLT for short. Or we can use data services if we are fine with a batch-based data acquisition process.

00:11:30 Then we also have SAP Replication Server. Now, for the non-SAP sources, we also haveother techniques.

00:11:37 Very often it's devices, like network devices, either wired or wireless. Then we have sensor ormachine data. And then we have social media or text data.

00:11:49 So, for those purposes, we very often use streaming technologies, stream technologies, for

example ESP, Event Stream Processer or SAP SQL Anywhere.00:12:02 And also, data services are very often used in this context. So bringing all this data into SAP

HANA can basically happen through all these tools.

00:12:11 Now, in the next unit, we're going to drill down into more details. We're going to talk about allthese different tools. Why they are different, and what they do differently, and what they'rereally good at.

00:12:21 So, with that, I'd like to conclude this unit. We talked about all the Big Data platformcomponents.

00:12:28 We compared in fact why SAP HANA is good for certain things, why Hadoop is the betterchoice for other reasons. And then there's SAP IQ.

00:12:41 We kind of talked briefly about all the acquisition techniques in the tools but, as I said before,

we are going to dedicate the whole next unit to that topic.00:12:53 Thank you for watching Unit 2 of this week, and I'll see you again in the next unit.


8/20

8

Week 2 Unit 3

00:00:12 Welcome back to the Driving Business Results with Big Data openSAP course. This is Unit 3of the second week, and in today's unit we're going to talk about the data acquisitiontechniques and technologies

00:00:23 as part of the SAP Big Data platform solutions. So our main challenge when we talk about theacquisition of data, especially Big Data, is that

00:00:35 for Big Data purposes, data is coming from all kinds of sources. So we're talking about datacoming from ERP systems, CRM systems, other enterprise systems whether SAP or non-SAP.

00:00:47 Then we have obviously unstructured data coming from social media, data coming from e-mails, weblogs. And also we're talking about signal data coming from sensors and machines.

00:00:59 All this data has to be somehow brought together and then put in as one foundation into ourmain platform. So, for the purpose of the data acquisition, we have various techniques andtools available.

00:01:13 We're going to be talking about them one by one, what their strengths are, and in which usecases they are the best, and what the differences between them are.

00:01:21 So, we'll be talking about solutions and tools, such as SAP Data Services or SAP System

Landscape Transformation, or SLT for short. Or we're going to talk about the streamingtechnologies, such as smart data streaming,

00:01:36 or also another tool called the SAP Event Stream Processor. And we're going to talk aboutrelatively new techniques, such as SAP smart data access,

00:01:46 and also SAP smart data integration. So first let's talk about SAP Data Services.

00:01:53 This is probably one of the most universal SAP data acquisition tools. It's been available for along time now. And it's a really good solution for data integration, data quality, data profiling,and data processing

00:02:09 regardless of the data source or the type of the data. So we can use it pretty much in all kindsof situations.

00:02:18 The capabilities of this tool are really powerful. It is capable of combining data from very, verydistinct data sources

00:02:28 and then integrating this data. So one of the main features is actually data integration here inData Services. We can integrate this data, so we can form a reliable data foundation for furtherprocessing.

00:02:41 It has very enhanced Hadoop support. And probably the most significant advantage is thatSAP Data Services is pretty much one server to execute all these capabilities.

00:02:55 It is also a design environment. So, in fact, in SAP Data Services we can do a lot of coding

00:03:03 through a visual graphical interface where we can actually transform the data, we can integratethe data coming from different sources

00:03:13 even if that data has a different structure. It also has an administrative console

00:03:20 to monitor and control all these activities. So, since this is a very powerful data integration tool,which is why we use this tool a lot

00:03:36 during our SAP Big Data acquisition processes. So very often we use the SAP Data Servicesfor acquiring social media data, unstructured text data

00:03:47 and also in other situations, which we will see in more detail in our SAP RDS Weeks 3, 4 and5, when we're going to refer to this tool very often.

00:04:00 Then we have our SAP System Landscape Transformation. From now on, I'm just going to saySLT for short. So SLT is a really old timer, an old player in this ecosystem.

00:04:10 It has basically been available since SAP R/3 It's been a couple of decades now.

00:04:19 And it's a really great solution to move real-time data between various systems, whether theyare in the same network, or sometimes in different networks.


9/20

9

00:04:31 And also it's capable of moving data into the cloud. It's main features are data replication.That's one.

00:04:39 The data replication feature of SLT is really powerful. It can do real-time, trigger-based datareplication. It can filter the data. We don't really have to replicate one-to-one, one databaseinto another database.

00:04:53 For that, we could just do database migration, for example. But here we are talking aboutdeltas.

00:05:01 So, once we load the data as an initial loading to our target platform, SAP HANA database inthis case, then we can actually schedule the trigger-based replication process, which onlyhandles the delta replication.

00:05:15 And that makes it, in fact, real-time and very, very high speed. Also, it's a transformation tool.

00:05:24 It's not as powerful, obviously, as SAP Data Services was, as we talked about in the previousslide. But still we have here some data transformation capabilities. With SLT, we can filter thedata,

00:05:37 for example, we want to bring in data just by certain key fields, IDs, and so on. Or we canactually sort of cut off parts of tables.

00:05:49 So, if a source table has hundreds of columns, we can decide that we don't need all that data

for our SAP Big Data processes. So we're just going to cut out certain fields and replicate oracquire just what we also need.

00:06:08 It is also capable of converting the data, data conversion techniques. So, what we can do hereis hide certain data but still get the data into our source.

00:06:17 With that, for example, we can make sensitive data anonymous. And the installation andoperation of this tool is actually very simple.

00:06:30 It's very easily deployable, in a matter of hours or maybe one day. It's really simple. And itbasically supports all the systems, all the applications

00:06:44 in an SAP environment. So whether it's an SAP Business Suite application or anything else, itcan support those replication sources.

00:06:57 Now, let's talk about streaming analytics, well the acquisition of data for streaming analytics.

Now, here we're talking about a very different approach.00:07:06 So, in the case of streaming analytics. We first of all talk about complex event processing,

which means multi-streams coming in.

00:07:18 It's very important to understand that this is very different from other replication techniques.Here we don't really store the data. So SAP Event Stream Processor is like a staging area.

00:07:29 Streams are coming in, multiple streams, from sensors, from machines, for example. And, thisdata is not stored anywhere.

00:07:38 This data is not necessarily going anywhere either. So we can build in algorithms in this tool tomonitor that data

00:07:49 and then create output streams only based on our rules. For example, if we monitor thetemperature of equipment, then we're receiving the streams constantly from the equipment orthe machine.

00:08:04 But ESP will monitor this and they will only create an output stream to us, to our visualizationtool, or SAP HANA,

00:08:15 only if, for example, the temperature is higher than a certain level that we specified in advance. Another very important feature of this tool is that it's capable of receiving multi-streams

00:08:31 and actually comparing those multi-streams. Let's use the same example of measuring thetemperature of equipment. We might specify that we want an alarm.

00:08:43 So that means we want to trigger an output stream from ESP only if the temperature is higherthan 90 degrees centigrade. But provided that also the ambient temperature is higher than,let's say, 35 degrees centigrade.


10/20

10

00:09:00 So that means that ESP can actually monitor those multiple streams at the same time andmake a smart decision and trigger the output stream only when we want it.

00:09:11 So that's why when we talk about the output stream being a capability of ESP, we're talkingabout basically an event-driven streaming process.

00:09:24 Now, another alternative to ESP is SAP HANA smart data streaming. This is just another tool,another technique to do streaming data, receiving streaming data.

00:09:38 It does the same job, this tool, but there are big differences. First of all, ESP is a standaloneproduct.

00:09:47 It can be installed and it can be utilized in a context of other situations that we don't even need,for example, SAP HANA as a recipient. We can just use a monitoring, visualization front end toreceive those output streams

00:10:04 and monitor those. But, obviously, very often we want the data to be stored in SAP HANA.

00:10:11 So in that case, we can do that with ESP by the way, but we can also utilize this tool the SAPHANA smart data streaming, which is actually installed by the SAP HANA installer. It's part ofthe SAP HANA installation in fact.

00:10:26 It can run on the same server. And, for example, we recommend that even if we decide to usesmart data streaming,

00:10:37 instead of ESP for development, or quality, or training purposes, we can actually install it onthe same database where SAP HANA is running, on the same host.

00:10:50 For productive purposes, obviously we recommend it to have its own host. As of SAP HANAsupport pack level 09,

00:10:59 we can use the SAP Event Stream Processor or HANA smart data streaming. And by now, asof Service Pack SP10, we can actually use them side by side as well.

00:11:15 Another advantage of this tool, smart data streaming, is that it has a so-called streaming studioplug- in, which goes into SAP HANA Studio.

00:11:26 So all the monitoring and administrative tasks are happening in the same interface. It is veryeasily deployed.

00:11:37 It can be deployed in the cloud, for example, so there are other benefits of this tool versus the

ESP tool. Now the ESP tool is more powerful, more capable. Also we can use it in differentcontexts, in different environments.

00:11:53 Now let's talk about another technique. This is called the SAP HANA smart data access.

00:11:59 So, this is relatively new and is available as of SAP HANA SP Support Pack Stack 06. And,with this, we can actually expose data from remote sources as virtual tables in SAP HANA.

00:12:14 Now this gives us a really great tool to integrate virtual tables with physical tables in SAPHANA without the need to bring the data, which is basically virtual tables, from the remotesources.

00:12:31 So we can actually access the remote data as local data. The advantage is that we canleverage the processing capabilities of those remote sources.

00:12:43 We can push down the processing into those. Instead of bringing the data into SAP HANA.

00:12:48 And, also on the other hand, if let's say, those remote sources are not capable of doing certainthings that we need to do or want to do in HANA,

00:12:55 then this technique compensates for the lack of those capabilities because we can use HANAto process remote data as local data.

00:13:07 And then we have our last tool, it's SAP HANA smart data integration. Now this is really new.It's available as or SAP HANA SP09.

00:13:16 And this is kind of like a Swiss knife. It has all the capabilities. It can do everything in one. So,obviously, if we compare it with the other tools, it might not be as powerful as SAP DataServices.

00:13:28 It might not be as flexible with traditional ERP systems, like SAP SLT. But the advantage ofthis is, many customers might like to use this because it can do all in one.


11/20

11

00:13:44 It can support both batch and real-time acquisition processes. It can also do certain datatransformations both on batch and real-time data.

00:13:57 And it can integrate data between on-premise and cloud systems. It has one connectivitytechnique to connect to all the data sources

00:14:13 and also has one user interface to do all these administrative and operational tasks that weneed to do here. So this was our last tool for data acquisition.

00:14:27 So, with that, we've covered all the available tools. Thank you for watching this video. This wasour Unit 3 when we talked about data acquisition into our SAP HANA Big Data platform.

00:14:41 In the next unit, we're going to talk about the applications and analytics that we use to analyzeand also act upon the Big Data. Thank you for watching me in this video, and I'll see you in thenext unit.


12/20

12

Week 2 Unit 4

00:00:12 Welcome back to the Driving Business Results with Big Data openSAP course. So this ourUnit 4 of the second week of the course, and we're going to be talking about SAP Big Dataapplications and analytics.

00:00:24 So, in the previous units we covered one of the major components of the SAP Big Datasolutions or SAP HANA Big Data platform. And in the next unit, Unit 3, we saw all the

acquisition technologies and techniques

00:00:39 that we use to get the Big Data into our platform. So today we're going to talk about all theSAP Big Data solutions,

00:00:47 applications, and also a bit about analytics. We already saw the use cases that we are going torefer to again today.

00:00:59 So, there is a lot of use cases where Big Data resides and can be extracted and then thevalues and real results found in that Big Data. At the moment, we have a whole list of SAP BigData applications and solutions

00:01:14 that we consider as Big Data. So those help us to solve Big Data problems. We're going to talkabout those now in detail. I'm not going to list them here on this screen.

00:01:25 But we can see them in the subsequent slides where we're going to be talking about them one

by one. So, let's start with out SAP hybris marketing.

00:01:36 SAP hybris marketing sounds like something new, a relatively new product. And indeed it waslaunched in February of this year.

00:01:45 But, in fact, this product, this solution is built on the foundation of two existing products. One ofthem is SAP Customer Engagement Intelligence,

00:01:54 which brings a sophisticated customer profiling and segmentation of those customers basedon data coming from CRM systems or other like enterprise systems.

00:02:06 And then we have another component which came into this picture. It's SeeWhy, anothersoftware acquired by SAP last year.

00:02:16 And this gives us real-time buying signal detection, and also one-to-one marketing. So, reallythe differentiator between traditional marketing techniques

00:02:29 versus what we can bring now with SAP hybris marketing through the technology, SeeWhysoftware is basically personalized one-on-one marketing campaigns.

00:02:42 So no more mass e-mailing and mass phone calls to customers with the same message. Wecan detect, using social media feeds and analysis,

00:02:53 what individuals want and what they think about our products so we can target themindividually. Now, I'm not going to spend any more time on this product because we're going todedicate a whole week

00:03:04 to SAP hybris marketing. It's going to be our last week, Week 5, of the course. And, as we'vealready mentioned before, we're also going to have a guest speaker from the SolutionManagement team.

00:03:18 Our next Big Data solution is SAP Demand Signal Management. Now, this is designed for the

retail industry.

00:03:27 There are a lot of challenges. Probably the first challenge is that the retail point-of-salelocations generate an enormous amount of data.

00:03:39 Millions of transactions on a daily basis, and a lot of data coming in. So, two main problemsthat we are trying to address with this solution.

00:03:48 The first is how to measure the promotion effectiveness. Today, the retail industry spends over300 billion dollars on sales and marketing promotions and campaigns.

00:03:59 Now, how can we know if they are actually effective? How can we know if the money spentactually brings results? According to experts and even the insiders of the retail industry, it's50% art and guess

00:04:14 and 50% scientific approach. So, our goal here, with this application, is to bring that 50%scientific approach closer to


13/20

13

00:04:23 Maybe not completely 100%, but very close to it. Another challenge that the retail industry isfacing is how accurately can they forecast the future demand.

00:04:35 So, one of the first situations in retail is when there's a demand for something and there'sinsufficient stock to meet that demand.

00:04:43 So, obviously, missed demand means lost sales opportunities. With SAP Demand SignalManagement, we offer some answers to these challenges and questions.

00:04:55 So, first of all, we can bring in a massive amount of POS data, point-of-sales data into ourplatform, and analyze that data.

00:05:04 Then we can also rely on data coming from unstructured data sources, like social media, forexample. This solution, this product, comes with pre- packaged, pre-built role-baseddashboards and

00:05:18 analytic reports. And it also integrates very well with other components of the SAP ecosystemlike

00:05:26 for example, SAP Trade Promotion Management, and CRM itself. So, as a result, we canreally gain a real-time

00:05:37 or we can sense real-time actual demand signals from the market. And also, we can quicklyspot those deviations in trends, in demands.

00:05:46 So we can have a much more accurate forecasting of the customer demand.00:05:54 Our next application is SAP Fraud Management. Now, this is tailored to banking or the

insurance industry.

00:06:01 And we really have three main objectives here in our sight. First, we want to detect fraudearlier.

00:06:10 We don't want to wait until fraud happens. We want to detect it and react in time to avoidfinancial loss.

00:06:18 Another main goal is to improve the accuracy of the fraud detection. Now, how annoying is itwhen our bank calls us and tells us that basically our bank or credit card

00:06:30 has been canceled because there was a fraudulent activity or at least they suspect there was. And then they ask us all about our activities to know if we made those purchases

00:06:41 and we say, yes, it was all me, I did that. And then basically, sorry - false alarm. So, then Ihave to wait for the mail until I get my new credit card or debit card.

00:06:53 Another very important objective of this solution is to basically prevent fraud. So how do we dothat? We're trying to build predictive models

00:07:05 to predict the situation in which there is a possibility of fraud, how fraud happens. So, basically,this fraud management application comes with a feature where we can design,

00:07:16 model, basically, situations based on historic data. We analyze the historic data, we see thosepatterns, for example, how was credit card usage

00:07:27 happening, what kind of activities were there, and what's the first sign that there might befraud. After that, we can also do a fine-tuning.

00:07:38 There's a so-called fraud calibrator. So we can really fine-tune and set up the exact model that

we need.00:07:47 And after that, we run the model, we run our fraud detection. And, as a result, we can start an

investigation.

00:07:53 Now, this actually happens before any fraud occurs. We are looking proactively and we aretrying to predict what's going to happen.

00:08:02 This application also comes with a whole set of reports and dashboards for all these activities.Then, we have the next application.

00:08:11 Now, back to the retail industry. So we have our SAP Sales Insights for Retail. Now, this isreally a toolbox.

00:08:19 So, there are different tools in this application, and all these tools are designed to betteranalyze promotion effectiveness as well, promotion planning


14/20

14

00:08:30 and also assortment optimization. What is the right assortment in a certain retail location. Andalso managing stores and product groups, so we can monitor how they perform.

00:08:40 There are three main components of this application. As we've said, it's kind of a toolbox. Solet's say we have separate tools.

00:08:47 First we have our value driver tree. Now, the value driver tree is really a quick guided root-cause analysis for performance changes.

00:08:57 So, we're looking at performance indicators, KPIs, and we detect changes to the trend. Andthen we're trying to look at other performance indicators, other KPIs, to see how they're relatedto each other.

00:09:10 Maybe a change happens in one indicator because of a change in another indicator. So it'skind of like trying to find a correlation between different KPIs.

00:09:21 Then we also have a tool called affinity insight. Affinity insight is basically a market basketanalysis.

00:09:28 We're looking at, literally, what customers put in their shopping basket. So, for example, wecompare shopping baskets. We look at what product group A is sold most frequently with.

00:09:43 What is the other product group that is sold with product A, for example. We also look at theprobability of, if the customer buys product A,

00:09:52 then what's the probability that they will also buy product B. Now this is a very useful tool. Avery good use case in this example would be promotions.

00:10:02 Very often the retail industry wants to run a promotion on a product to actually improve thesales of another product. So, we can analyze that typically a customer would most probablybuy that product

00:10:17 if there was a discount given to another product. So, we can analyze and find thosecorrelations again.

00:10:25 The other tool is the key item list. The key item list is really kind of an analysis of all the majorkey figures

00:10:32 like, for example, the most critical article numbers or the best-selling product groups, or howwell the different retail locations perform. Our next SAP Big Data application is SAP PredictiveMaintenance and Service.

00:10:49 Now, this is really in the area of Internet of Things. This is one of those applications that fitsinto this whole framework we have.

00:10:59 This is basically a cloud-based, cloud-deployed offering by SAP. It's deployed and actuallydeveloped, built, in SAP HANA cloud platform.

00:11:12 Obviously, the main goals of this application are to provide 24/7 monitoring of equipment andassets based on monitoring the signal data, machine data as well.

00:11:25 And then to predict the possible malfunction and breakdown of that equipment. So we're tryingto prevent a breakdown. We want to proactively do a repair service if we see that

00:11:36 there is a trend that machinery or equipment might break down. So, with that, our goal is tooptimize the maintenance and service operations of an organization.

00:11:50 And, finally, let's look at another very exciting and relatively new topic. It's back to the internetof sales and Industry

00:11:59 So here we have a whole line of different capabilities. So this is really not one application thatwe're looking at, connected logistics and connected manufacturing,

00:12:08 but it's really like a suite of different applications that all belong to the same idea. Now, what isthat idea really? The whole thing starts with our embedded systems.

00:12:22 Let's take a step back and move away from manufacturing or logistics. If you just look at oureveryday life: Today we all walk around with smartphones and all kinds of other devices thatcan measure data.

00:12:35 There's also sensors built in in our home appliances, our refrigerator, our washing machineand all those appliances. Our computers constantly provide data about what we do,


15/20

15

00:12:47 what we look at, what we browse. So, the idea is that we can measure pretty much everythingabout ourselves.

00:12:55 Now, if we add mobile communication to that, that data can be transmitted and collectedobviously. And, then we add us, humans, to this ecosystem.

00:13:07 So, the devices, the sensors, the machines, and us humans. We're talking about a kind ofcyber- physical system that we can basically call Internet of Things.

00:13:18 Now, let's take this whole idea and apply it to manufacturing. So let's take all these smartdevices and smart equipment, and smart production tools

00:13:29 - shopfloor, which we can combine and call a smart factory. Now, in the smart factory, all thesemachines can communicate with each other.

00:13:38 They can find out where there's free capacity, so they can shift capacity to where it's moreavailable. At this point, we're already talking about Industry 4.0.

00:13:50 And there're also the products that we produce through smart factories. We product let's callthem smart products, smart items, because they can actually tell the machine how they wantto be manufactured.

00:14:03 They can measure their own signals. And if those signals are not within the manufacturingstandard, they can tell that to the machine

00:14:13 and all kinds of parameters can be changed automatically. Now let's add humans to that, theoperators of all these machines.

00:14:21 Those operators can walk around on the shop floor with the smart devices, smart glasses. So,we're really talking about something that's an augmented reality. We're talking aboutaugmented operators.

00:14:33 And as those operators, those humans, communicate with each other, it's becoming like asocial network. All these people, they work together, they communicate together, real-time,non-stop.

00:14:45 So, let's go back now to our applications here, or actually suite of applications. We have awhole line of products which support this whole idea.

00:14:54 We have overall equipment effectiveness, energy management, process optimization, just toname some of them. And then we have the connected logistics

00:15:04 where this whole idea from the shopfloor also then applies to the shipping process, thelogistics process. Previously, maybe trucks for a long time had GPS.

00:15:16 They already had the coordinates and knew when they would approximately arrive at theirdestination. But we can do much more than that. We can actually predict the estimated time ofarrival

00:15:30 if we also add to the equation data like traffic reports. We can add data on weather patterns,forecasts, and so on.

00:15:38 And suddenly we're talking about much better prediction of those logistically significanttimelines. So, this was our not complete but almost fullest of available SAP applications thatsupport Big Data.

00:15:57 Obviously, there's also analytics, which we talked about again. But we thought it was importantto include it in this unit. Again, analytical applications and tools that help us to engage,visualize, and predict.

00:16:13 I'm not going to go into detail about this. We're going to, as I mentioned before, spend muchmore time on this in the coming weeks

00:16:22 when we talk about the SAP rapid-deployment solutions. So, with that, I'd like to close Unit 4 ofthis week.

00:16:29 We talked about SAP Big Data applications and analytics. In the next unit, we're going to lookat Big Data science and Big Data services.

00:16:39 So, we're going to look at what is really the difference between traditional analytics and datascience. Why is it more, and what can it help us to achieve?

00:16:50 Thank you for watching this video, and I'll see you again in the next unit.


16/20

16

Week 2 Unit 5

00:00:11 Welcome back to the last unit, Unit 5, of the second week, and it's still the Driving BusinessResults with Big Data openSAP course. So, we talked in the previous units about the first twocomponents of SAP Big Data solutions.

00:00:25 We talked about the SAP HANA platform for Big Data, and we also talked about theacquisition techniques and the applications and analytics around SAP Big Data solutions.

00:00:36 So, today we're going to talk about the last component, which I think is one of the most excitingcomponents because this is where we really find out how to find the patterns, how to find thosecorrelations

00:00:49 and those hidden secrets, basically the results behind Big Data. How we apply it to ourbusiness. So, first of all, we've been doing analytics for ages now.

00:01:03 But let's distinguish, let's make a difference between standard, traditional analytics and datasciences because these are different things.

00:01:12 So, what is really the difference? Let's identify data science as transforming data into businessresults, business value using mathematics and algorithms. A scientific approach.

00:01:26 Compared to that, traditional analytics is based on the standard ways of looking at data suchas standard reporting, drill-downs, filtering. So really not doing too much with that data.

00:01:41 Just looking at the data as is. So, obviously beyond the standard reporting and ad-hocanalysis, there is different levels

00:01:51 of involvement of the scientific approach to data analysis. We have data mining first.

00:01:59 Data mining is probably the most simplified, the easiest way of looking at data with a scientificapproach. So what do we do here? Here we use mathematic standards

00:02:13 to look at the data. So we make some changes to the data, for example, we do clustering ordecision trees. We apply regression algorithms. But we're really not modifying too much.

00:02:26 We're not really changing the data too much. So we do that with the purpose of findingpatterns.

00:02:33 Now very often this is not sufficient, very often we just do all these techniques applying thosestandard mathematical algorithms, but we still don't find what we are looking for.

00:02:43 It seems that there is no result. So, that's the point when we go to the next level, and we try toapply models.

00:02:51 We create basically mathematical models, algorithms, which are pretty much formulas. Withthose formulas, we're trying to actually look into parameters.

00:03:04 For example, we try to see if we change one parameter, how does it affect the resulting KPI.Then we take another parameter, we change that, we see how that then changes.

00:03:17 So there we can find out details that were previously unknown to us. And then we have dataoptimization.

00:03:26 So, imagine that with modeling, for example, we reached a level of sophistication where we

know exactly that parameter A, B, C, and D, changing one way or the other

00:03:37 will positively affect our profitability margin, for example. But, now, the question still is which isthe best way to do that?

00:03:46 That's the point where we have to create and apply optimization algorithms to find out which isreally the best way. With that, we're trying to find the best answer to a solution.

00:03:58 So that we can apply it to our business reality, our business operations. So let's look at someexamples.

00:04:06 We're going to start with the easy, simple data mining. Let's take an example from the retailindustry.

00:04:13 Let's imagine a retail company, a consumer product company, decides to launch two newflavors of a soft drink. And then after six weeks, they look at the data, sales figures, and they

see that product A sells better than product B.


17/20

17

00:04:30 Well, that's a great result, so we can make a conclusion that product A is the one that we wantto go with. And, for B, maybe we want to give it some promotions, some discounts, so we canactually increase the sales of that product.

00:04:42 Now that could be a critical mistake. Because if we look at the same data from a differentangle,

00:04:50 for example, we look at repeat purchases, how many times that product B was actually sold,the same flavor, versus product A

00:05:01 and we realize that actually product A had a higher sales volume like numbers in that first sixweeks. But that was maybe because it was a new flavor.

00:05:13 It was more preferred by customers, they liked the color of the packaging better, for example. And product B was not so favored.

00:05:22 But once they bought both products and they compared, they tasted it actually, they realizedthat they preferred product B, and they started to buy that more often than product A.

00:05:32 So now we can make a real conclusion that actually product B is the one that is moresuccessful. So, this is a good example of data mining.

00:05:43 Let's look at another example. Now this is getting more complex here because we're looking atsupply chain optimization. This is very critical, for example, for businesses where there is ahuge spare part inventory needed

00:05:56 and there's a huge supply chain process behind that spare parts supply. So those industriescould be the airline industry, shipping industry, transportation industry.

00:06:08 Everywhere there is very complex machinery or equipment at hand, like airplanes or trucks, orwhatever. In this case, we are looking at an example from a railway company.

00:06:21 So, obviously, railway engines, they have very complex, spare parts distribution behind therepair and maintenance organizations. Typically, in the traditional way, in the classical world,they set up in the ERP system a reorder point.

00:06:42 So, as soon as the inventory of that spare part reaches that reorder point, it triggers thereplenishment process and repurchasing that spare part, putting it on the stock.

00:06:51 Now is this the best way to do this? Probably not. So why not look at historic data?

00:06:56 Why not analyze how the supply was affected by also other factors, like travel volume, forexample, weather patterns, maintenance records, age of the equipment.

00:07:13 So, if we look at that, then based on that historic data, we could build up models to forecast thefuture demand on the spare parts.

00:07:23 And, we can suddenly realize that sometimes in some repair/maintenance locations thatreorder point is actually lower than in the others.

00:07:36 Sometimes, we'll realize that in some locations we never really reached that reordering point,but still because we applied the same numbers to all the locations we were reordering

00:07:47 and having an extra inventory which was not necessary. So, with that, with predictiveforecasting, we can really optimize the supply level.

00:07:56 And, also, we can apply different reorder points, different stock levels, to the different

locations, considering all the different factors.00:08:07 Let's look at another example. This is also sort of a data modeling situation, media newspaper

publishing. Here the publishers have a very difficult task at hand.

00:08:23 They have to predict, they have to guess how many copies they have to distribute to thedifferent retail POS locations.

00:08:32 Now, the POS locations are very different. Some newsstands sell more than others. Somenewsstands have traffic that fluctuates.

00:08:42 Some newsstands have days or weeks when they sell the minimum, and other days becauseof special events when they sell much more. So, how to take all this data and consider all this?

00:08:53 Again, with historic data and building predictive and forecasting models, we can actually

modify what our safety stock is, which we add to that predicted sales volume.


18/20

18

00:09:04 And, with that, we can actually apply the model and change by location. So, we can see thatlocation A actually needs a higher safety stock

00:09:16 because they sometimes, in some time periods, they actually sell more than the other location,while the other location has sometimes days when they don't sell anything.

00:09:28 So we can vary those supplies that we send out to those POS locations. And one of the otherinteresting points is that all we did here was play with the numbers.

00:09:43 We analyzed the numbers, and based on historic data, and based on information comingmaybe from traffic forecasts for a specific airport,

00:09:54 or railway station, we predict, we forecast that certain dates, certain events will generate moresales. We didn't really change the business process.

00:10:04 We're still printing the newspapers the same way, delivering to the POS location and sellingthe same way. Just by applying some numbers, we actually optimized our distribution model.

00:10:16 Another example would be the utilities industry. In utilities, there are lots of ways to actuallywork with Big Data and find values in Big Data.

00:10:29 So, in this example, we're not going to talk about maintenance, we're not going to talk aboutsensor data. And then based on that sensor data, predictive maintenance, for example,

00:10:41 asset services, and so on. We're talking about another aspect of sensor data. So when we

have a power grid,00:10:49 there are millions of sensors in a power grid. And those sensors are constantly pushing that

data, streaming that data into our engine. And what happens is that very often the data isincomplete.

00:11:01 Or very often the data is faulty. Sometimes, the measurement is really out of whack, it reallyhas nothing to do with reality.

00:11:08 So what happens then? We have to actually go and repair a sensor that could be just aphysical activity.

00:11:16 Or we can actually, based on business events, business rules, we can, let's say, calculatesomething. Now why not apply a different method? Why not utilize data sciences

00:11:30 and basically fill out those missing values with optimization algorithms looking at historic data

coming from that sensor, looking at other parameters coming from other sensors00:11:43 related to, for example, weather, temperature, and so on. So, then what we do is, we basically

calculate what is the most possible value

00:11:52 that should have or could have come from that sensor if everything had been ok. And what weget is complete data.

00:12:00 So, basically, applying data sciences and optimization algorithms here, we can fill those gapsand make data quality better. So, these were a couple of examples, use cases, like how datasciences can apply to real-life examples.

00:12:20 Now, who really does that? Here at SAP, we have a data science service organization.

00:12:25 So, this is basically an organization that has 3 global teams located in different regions inEurope and in North America.

00:12:37 And they have extensive experience in mathematical modeling, forecasting, simulations,optimizations. And this is a team of PhD-level experts who actually do all this research and allthese consulting services for our customers.

00:12:55 There's also a methodology involved. So, it's well proven. We call it our Big Data servicescustomer journey

00:13:05 where basically the service organization can provide a step-by-step roadmap to the customers. And the good thing is that customers can actually join this roadmap at any stage.

00:13:16 So, depending on their maturity level on Big Data, they can, for example, start with theinnovation session. Sometimes customers understand that they have a Big Data problem.

00:13:26 But they're kind of hesitant to really get started because they don't know where to start. So, inthat situation the SAP data science services organization offers an innovation session.


19/20

19

00:13:38 This is basically a design thinking workshop where they look into what's going on, what thecustomer would like to get out of Big Data, what kind of values, real results,

00:13:47 and then there's going to be a foundation for further work. And then we can have a Big Dataadvisory process.

00:13:56 This is basically on the strategy level where customers together with our service organizationcan work out the whole Big Data strategy. How to get from A to Z, and after that, Big Dataproof of concept,

00:14:10 which then goes down into details when the whole project plan and all details of this wholeplan are worked out. Architecture deep dive, and then the Big Data project can be started andrealized.

00:14:23 Also, there are sessions about data discovery and insight, and also enablement sessions. So,as we can see, SAP really has a full picture about Big Data solutions.

00:14:36 We talked about things like the SAP HANA platform. So, the technology at hand. We alsotalked about all the techniques and tools which help us to get that Big Data.

00:14:45 Then we talked about the applications and analytics. So, the great thing about this is that SAPdoesn't only have

00:14:55 the platform and the technology to get the data, but we can actually operationalize this data.We can actually act upon the data.

00:15:04 And we also have the know-how, the skills to provide the services around this. So, with that,we can conclude our Week 2, which we dedicated to SAP Big Data solutions.

00:15:17 And in the next weeks, as we promised, we're going to get down really under the hood. We'regoing to open the engine and we'll see what's inside and how to actually work with this.

00:15:29 So, we're going to be talking about three different rapid-deployment solutions. In the nextweek, we'll start with the SAP Big Data Intelligence RDS.

00:15:38 I'm looking forward to seeing you again. All the best till then, thank you.


20/20

www sap com

© 2015 SAP SE or an SAP affiliate company. All rights reserved.No part of this publication may be reproduced or transmitted in any formor for any purpose without the express permission of SAP SE or an SAPaffiliate company.SAP and other SAP products and services mentioned herein as well as theirrespective logos are trademarks or registered trademarks of SAP SE (or anSAP affiliate company) in Germany and other countries. Please seehttp://www.sap.com/corporate-en/legal/copyright/index.epx#trademark

foradditional trademark information and notices. Some software productsmarketed by SAP SE and its distributors contain proprietary softwarecomponents of other software vendors.National product specifications may vary.These materials are provided by SAP SE or an SAP affiliate company forinformational purposes only, without representation or warranty of any kind,and SAP SE or its affiliated companies shall not be liable for errors oromissions with respect to the materials. The only warranties for SAP SE orSAP affiliate company products and services are those that are set forth inthe express warranty statements accompanying such products and services,if any. Nothing herein should be construed as constituting an additionalwarranty.In particular, SAP SE or its affiliated companies have no obligation to pursueany course of business outlined in this document or any related presentation,or to develop or release any functionality mentioned therein. This document,or any related presentation, and SAP SE’s or its affiliated companies’

strategy and possible future developments, products, and/or platformdirections and functionality are all subject to change and may be changed bySAP SE or its affiliated companies at any time for any reason without notice.The information in this document is not a commitment, promise, or legalobligation to deliver any material, code, or functionality. All forward-lookingstatements are subject to various risks and uncertainties that could causeactual results to differ materially from expectations. Readers are cautionednot to place undue reliance on these forward-looking statements, whichspeak only as of their dates, and they should not be relied upon in making
http://www.sap.com/corporate-en/legal/copyright/index.epx#trademarkhttp://www.sap.com/corporate-en/legal/copyright/index.epx#trademarkhttp://www.sap.com/corporate-en/legal/copyright/index.epx#trademark

Documents

OpenSAP Big1 Week 2 Transcript