europe 2016 - bi-podium.nlbi-podium.nl/mediafiles/upload/dmzone16/dmz2016_program.pdf · the...

18
10 th & 11 th OCTOBER 2016 ABION Spreebogen Waterside Hotel Berlin, Germany Version: 2016-05-02 EUROPE 2016

Upload: others

Post on 22-May-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

10th & 11th OCTOBER 2016ABION Spreebogen Waterside Hotel

Berlin, GermanyVersion: 2016-05-02

EUROPE2016

Monday, October 10th

7:00 – 9:00 Breakfast and Registration

Agile as a complex ecology:from manufacturing toservice

Dave SnowdenCognitive Edge[United Kingdom]

Framing Business Problemsas Data Mining Problems

Asoka DiggsIntel Corp.[United States]

Welcome and Announcements

Lunch

KEYNOTE: Humans are neither ants nor data processorsDave Snowden, Cognitive Edge [United Kingdom]

Business InformationModeling: A Methodology forData-Intensive Projects, DataScience and Big DataGovernance

Torsten PriebeSimplity[Austria]

Customer case ofHeureka/Erwin

PS-3C Modeling: proposedway of agile EDW datamodeling on the Data Lake

Rogier WerschkullDIKW[Netherlands]

Customer case of Teradata A Great Challenge: DesigningA New DWH For A Bank

Murat CetinkayaING Bank[Turkey]

Data Modeling for NoTables

Thomas FrisendalTF Informatik[Denmark]

9:00 – 12:00

12:00 – 13:15

13:15 – 13:30

13:30 – 14:30

14:30 – 15:30

15:30 – 16:00

16:00 – 17:00

Foundational DataModeling

Tools & NoSQL Advanced DataModeling

Communication Skills& Case Studies

Business InformationModeling using the fact-based approach

Clifford HeathInfinuendo[Australia]

Advanced Data ModelingChallenges Workshop

Steve HobermannSteve Hoberman& Associates, LLC[United States]

Modeling of ReferenceSchemes

Dr. Terry HalpinProfessor at INTI InternationalUniversity, Malaysia[Australia]

CDMPCertification

Testing

Testing

Testing

About Ontologies – theFoundation of Data Modeling

Peer M. CarlsonDörffler & Partner[Germany]

Afternoon Snacks

GR GL PA CH

GR CH PA GL

PA GL GR CH

Rooms: GR|Grunewald GL|Glienicke PA|Pankow CH|Charlottenburg KÖ|Köpenick

DMZ on Twitter:#DMZone

Get the Guidebook Appand schedule your DMZ:

www.guidebook.com/getitRedeem Code: dmzeurope

NoSQL Data ModelingOverview

Steve HobermannSteve Hoberman& Associates, LLC[United States]

Fact-Based ModelingWorkshop

Dr. Terry HalpinProfessor at INTI,Int. University, Malaysia[Australia]

How to compose a fact-basedmodel into any kind ofschema

Clifford HeathInfinuendo[Australia]

From Conceptual to PhysicalData Vault Data Model

Dirk LernerITGAIN[Germany]

Data Modeling forSustainable Systems

Graham Witt[Australia]

9:00 – 12:00

Welcome and Announcements

KEYNOTE: From BI to AI - Models, Ethics and EconomicsBarry Devlin, 9sight Consulting [South Africa]

12:00 – 13:15

13:15 – 13:30

13:30 – 14:30

Why Business Analysts andData Modelers should holdhands more

George McGeachieMetadata Matters[United Kingdom]

Data Modeling forPolyglot Persistence

Phill RadleyBT[United Kingdom]

The Journey from ER Modelerto Data Scientist

Asoka DiggsIntel Corp.[United States]

smartDATA and Industrie 4.0

Dr. Siegmund Priglingerdr.priglinger consulting GmbH[Germany]

14:30 – 15:30

15:30 – 16:00 Afternoon Snacks

Building a Data Vault usingan Analytical Database

Johan van der KooijDe Bijenkorf[Germany]

Creating Ontologies fromyour Existing E/R Data Models

Nicolas LevequeDeutsche Bank[United States]

Positioning Data Modeling inthe Big Data Area

Martijn ImrichXomnia[Netherlands]

16:00 – 17:00

Tuesday, October 11th

7:00 – 9:00 Breakfast and Registration

Foundational DataModeling

Tools & NoSQL Advanced DataModeling

Communication Skills& Case Studies

CDMPCertification

Testing

Testing

Testing

Rooms: GR|Grunewald GL|Glienicke PA|Pankow CH|Charlottenburg KÖ|Köpenick

GL GR CH PA

GR GL PA CH

GL CHPAGR

8:00 – 8:30MorningSession

Casetalk –Data Modeling by example

Marco WobbenBCP Software[Netherlands] PA

Lunch

4

DMZ Europe 2016

Humans are neither antsnor data processorsDave Snowden, Cognitive Edge

Murray Gell-Mann is an American physicist whoreceived the 1969 Nobel Prize in physics fa-mously said : "The only valid model of a humansystem is the system itself”. In this address Prof.Snowden, who is Director of the Centre for App-lied Complexity at Bangor University in Wales,and Chief Scientific Officer of Cognitive Edge willoutline new approaches to the use of distributedethnography in understanding requirementscapture and knowledge-information flows in or-ganisations. Focusing on allow unarticulatedneeds to surface the approach offers a radicalnew way to model both cultural and informationaspects of an organisation. Based on complexi-ty and cognitive science the approach seeks toscale understanding of both actual and potentialneeds in a fractal manner, self similar at all le-vels of management.

Dave Snowden is the foun-der and chief scientific offi-cer of Cognitive Edge. Hiswork is international in na-ture and covers govern-ment and industry lookingat complex issues relatingto strategy, organisational

decision making and decision making. He has pionee-red a science based approach to organisations dra-wing on anthropology, neuroscience and complexadaptive systems theory. He is a popular and passio-nate keynote speaker on a range of subjects, and iswell known for his pragmatic cynicism and iconocla-stic style.

He currently holds positions as extra-ordinary Profes-sor at the Universities of Pretoria and Stellenboschand has held similar positions at Hong Kong Polytech-nic University, Canberra Univerity, the University ofWarwick and The University of Surrey. He held theposition of senior fellow at the Institute of Defenceand Strategic Studies at Nanyang University and theCivil Service College in Singapore during a sabbaticalperiod in Nanyang.

His paper with Boone on Leadership was the coverarticle for the Harvard Business Review in November2007 and also won the Academy of Managementaware for the best practitioner paper in the same ye-ar. He has previously won a special award from the

Academy for originality in his work on knowledgemanagement. He is a editorial board member of se-veral academic and practitioner journals in the field ofknowledge management and is an Editor in Chief ofE:CO. In 2006 he was Director of the EPSRC (UK) re-search programme on emergence and in 2007 wasappointed to an NSF (US) review panel on complexityscience research.

He previously worked for IBM where he was a Direc-tor of the Institution for Knowledge Management andfounded the Cynefin Centre for Organisational Com-plexity; during that period he was selected by IBM asone of six on-demand thinkers for a world wide ad-vertising campaign. Prior to that he worked in a ran-ge of strategic and management roles in the servicesector.

His company Cognitive Edge exists to integrate aca-demic thinking with practice in organisations throug-hout the world and operates on a network modelworking with Academics, Government, CommercialOrganisations, NGOs and Independent Consultants.He is also the main designer of the SenseMaker®software suite, originally developed in the field ofcounter terrorism and now being actively deployed inboth Government and Industry to handle issues ofimpact measurement, customer/employee insight,narrative based knowledge management, strategicforesight and risk management.

From BI to AI –Models, Ethics and EconomicsBarry Devlin, 9sight Consulting

The creative use of big data from social mediaand the Internet of Things is transformingbusiness models, empowering start-ups, andenhancing or destroying existing business value.Many proponents focus on marketing uses ofbig data, but most value—and disruption—willcome from its daily operational application innovel, transformative ways.

From slim beginnings thirty years ago in BI tosupport decision making, data collection andanalysis is now enabling AI systems to make fi-nancial recommendations or answer customercalls. They are diagnosing cancer and offeringtreatment options, and driving autonomoustrucks. Large swathes of decision making andaction taking are being automated, untouchedby human hand. Even complex decisions are in-

5

DMZ Europe 2016

creasingly augmented by cognitive computingsystems.

Making the right design choices is vital in desi-gning and managing such environments. Howe-ver, true business and IT leadership requiresconsidering the broader ethical and economic is-sues raised by this transformation. Concernsaround personal privacy, employment and socialdisruption must all be urgently addressed if indi-vidual businesses and society at large are tosuccessfully navigate this data driven transfor-mation of all aspects of business and technolo-gy, and even society.

Dr. Barry Devlin is amongthe foremost authorities onbusiness insight and one ofthe founders of data ware-housing, having publishedthe first architectural paperon the topic in 1988. Withover 30 years of IT experi-

ence, including 20 years with IBM as a DistinguishedEngineer, he is a widely respected analyst, consultant,lecturer and author of the seminal book, Data Ware-house—from Architecture to Implementation, and nu-merous white papers, blogs and more. His new bookBusiness unIntelligence—Insight and Innovationbeyond Analytics and Big Data was published in2013. Barry is founder and principal of 9sight Consul-ting. He specializes in the human, organizational andIT implications of deep business insight solutions inall technology environments. Barry is based in CapeTown, South Africa and operates worldwide.

NoSQL Data Modeling OverviewSteve Hoberman,Steve Hoberman & Associate

NoSQL implementations are often built with litt-le or no data modeling…or at the other extremecompletely over-architected – both ends of thespectrum producing suboptimal results. In thispresentation, understand how data modelingcontributes to the process of learning about thedata, and is therefore a required technique,even when the resulting database is not relatio-nal. We will explore what distinguishes NoSQLfrom a relational database, and what distinguis-hes a relational database from relational mode-ling. We will cover the optimal situations each

of the four types of NoSQL databases can beused. We will learn a streamlined approach todata modeling that ensures the NoSQL databasewill be built with a thorough understanding ofthe current requirements, as well as designed tohandle future needs. We will cover conceptual,logical, and physical data modeling for NoSQL,and how each stage increases our knowledge ofthe data and reduces assumptions and poor de-sign decisions. Exercises will be used to reinfor-ce techniques.

Steve Hoberman has trai-ned more than 10,000people in data modelingsince 1992. Steve is knownfor his entertaining and in-teractive teaching style(watch out for flying can-dy!), and organizationsaround the globe have brought Steve in to teach hisData Modeling Master Class, which is recognized asthe most comprehensive data modeling course in theindustry. Steve is the author of nine books on datamodeling, including the bestseller Data ModelingMade Simple. One of Steve’s frequent data modelingconsulting assignments is to review data modelsusing his Data Model Scorecard® technique. He is thefounder of the Design Challenges group and recipientof the 2012 Data Administration Management Asso-ciation (DAMA) International Professional Achieve-ment Award.

Agile as a complex ecology: frommanufacturing to serviceDave Snowden, Cognitive Edge

The wider vision of the Agile manifesto too of-ten ends up as a series of structured methodsdesigned to produce a product based on an un-derlying manufacturing method. Indeed muchthat is Agile is drawn directly from a manu-facturing environment. Any manufacturingprocess a priori makes assumptions about thedegree to which something can be defined andmeasured. In an ecology some aspects aresusceptible to such a way of thinking but in ma-ny cases needs are not, indeed cannot be articu-lated given the speed of change of technologicalcapability. With the Internet of Things we aremoving to a more fragmented ecology in whichapplications will emerge from the three way in-

6

DMZ Europe 2016

teraction of people, software objects and thewider environment.

The underlying architecture and metaphor ofsoftware development will have to undergo ra-dical change to accommodate these changes.Understanding the difference between sequenti-al iteration in methods such as SCRUM and par-allel short cycle experimentation wheredefinition comes at the end, not the start of theprocess will be key. Techniques in which partialcopying creates rapid mutation of need to allownovel forms of capability to emerge are beingdeveloped based on metaphors drawn directlyfrom evolutionary biology, in particular exaptati-on (rapid repurposing of existing capability tomeet unanticipated needs) and epi-genetics(Culture activates or deactivates genetic capabi-lity) are starting to provide a fertile, more simple(but not simplistic) approach to scale than simp-ly extended the engineering metaphor.

This workshop will give participants early in-sight into both the theory and practice of thisapproach.

Dave Snowden is the foun-der and chief scientific offi-cer of Cognitive Edge. Hiswork is international in na-ture and covers governmentand industry looking atcomplex issues relating tostrategy, organisational de-

cision making and decision making. He has pioneereda science based approach to organisations drawingon anthropology, neuroscience and complex adaptivesystems theory. He is a popular and passionatekeynote speaker on a range of subjects, and is wellknown for his pragmatic cynicism and iconoclasticstyle.

He currently holds positions as extra-ordinary Profes-sor at the Universities of Pretoria and Stellenboschand has held similar positions at Hong Kong Polytech-nic University, Canberra Univerity, the University ofWarwick and The University of Surrey. He held theposition of senior fellow at the Institute of Defenceand Strategic Studies at Nanyang University and theCivil Service College in Singapore during a sabbaticalperiod in Nanyang.

His paper with Boone on Leadership was the coverarticle for the Harvard Business Review in November

2007 and also won the Academy of Managementaware for the best practitioner paper in the same ye-ar. He has previously won a special award from theAcademy for originality in his work on knowledgemanagement. He is a editorial board member of se-veral academic and practitioner journals in the field ofknowledge management and is an Editor in Chief ofE:CO. In 2006 he was Director of the EPSRC (UK) re-search programme on emergence and in 2007 wasappointed to an NSF (US) review panel on complexityscience research.

He previously worked for IBM where he was a Direc-tor of the Institution for Knowledge Management andfounded the Cynefin Centre for Organisational Com-plexity; during that period he was selected by IBM asone of six on-demand thinkers for a world wide ad-vertising campaign. Prior to that he worked in a ran-ge of strategic and management roles in the servicesector.

His company Cognitive Edge exists to integrate aca-demic thinking with practice in organisations throug-hout the world and operates on a network modelworking with Academics, Government, CommercialOrganisations, NGOs and Independent Consultants.He is also the main designer of the SenseMaker®software suite, originally developed in the field ofcounter terrorism and now being actively deployed inboth Government and Industry to handle issues ofimpact measurement, customer/employee insight,narrative based knowledge management, strategicforesight and risk management.

Framing Business Problems as DataMining ProblemsAsoka Diggs, Intel Corp.

This will be a brief introduction to some of theimportant concepts that arise in data mining ingeneral and classification analysis in particular.The focus in this presentation will be helping at-tendees to develop their ability to framebusiness problems as data mining problems, tounderstand some of the terminology and mentalmodels prevalent in data mining, and to beingable to recognize situations where data miningmight apply.

Through a demonstration driven presentationformat, attendees will be exposed to a variety ofdata mining topics. Throughout the demo, at-tention will be paid to points of intersection

7

DMZ Europe 2016

with traditional data modeler work products in-cluding the data dictionary and dimensional da-ta model. I will also leave you with a variety oflinks and references for additional study on yourown.

Particular topics that will be addressed:• Formulating business problems as data

mining problems• Supervised Learning• Classification methods including Decision

Tree and Logistic Regression• Model Validation• No math!

After 15+ years of experi-ence in a variety of datamanagement disciplines,including database admi-nistration, ER modeling,ETL development, and dataarchitecture, I am transitio-ning into the world of ana-

lytics / data science. It turns out that there ISsomething even more fun than ER modeling, and I’vefound it. Today, my primary interest is in the organi-zational transformation involved in adopting analyticsas a source of competitive advantage. How does anorganization get there? What needs to be done? Whatorganization design and leadership changes are nee-ded to get the most benefit from analytics? I nowspend my work life practicing these new analytic mo-deling skills, and teaching others how to participateand contribute to predictive analytic projects.

Fact-Based Modeling WorkshopDr. Terry Halpin, Professor at INTIInternational University, Malaysia

Typically, analysis and modeling of an enterpriseor business domain is a collaborative processbetween the business expert, who best under-stands the business rules and requirements, andthe modeler whose main task is to elicit this un-derstanding and capture it in a formal modelthat can be executed (directly or indirectly viatransformation) by an information system. Ideal-ly, the model (including its representation of thebusiness rules) should be validated by thebusiness experts. Since business experts are of-ten non-technical, the model needs to be con-

ceptual in nature, cast in language that isintelligible to the business expert while stillbeing formal. Fact-Based Modeling aims to ad-dress this communication need in an optimalway.Fact-Based Modeling has been adopted by a fa-mily of related approaches, of which Object-RoleModeling (ORM) and SBVR (Semantics ofBusiness Rules and Business Vocabulary) arewell known examples. This presentation beginswith brief overview of fact- orientation in gene-ral and second-generation ORM (ORM 2) in par-ticular, highlighting its communication-basedmodeling procedure using controlled naturallanguage and visual support for conceptualizati-on of business rules. It explains ORM’s underly-ing principles, outlines recent improvements tothe methodology and associated tool support,and briefly contrasts fact-orientation with otherhigh level data modeling approaches (e.g. ERand UML). The bulk of the workshop then focu-ses on developing skills in constructing andtransforming data models by applying fact-ori-entation to practical examples of informationsystems, including advanced aspects, with theimplementation emphasis on relational databa-se systems.

Dr. Terry Halpin is a Rese-arch Fellow in ComputerScience at INTI Internatio-nal University (Malaysia),and a data modeling con-sultant. He previously heldsenior faculty positions incomputer science at TheUniversity of Queensland (Australia), and professor-ships in computer science at Neumont University(USA) and INTI International University. His prior in-dustrial experience includes many years in data mo-deling technology at Asymetrix Corporation,InfoModelers Inc., Visio Corporation, Microsoft Cor-poration and LogicBlox. His doctoral thesis formalizedObject-Role Modeling (ORM/NIAM), and his currentresearch focuses on conceptual modeling and rule-based technology. He has authored over 200 technicalpublications and nine books, and has co-edited ninebooks on information systems modeling research. Heis a member of IFIP WG 8.1 (Information Systems), isan associate editor or reviewer for several academicjournals, is a regular columnist for the Business RulesJournal, and is a recipient of the DAMA InternationalAchievement Award for Education and the IFIP Out-standing Service Award.

8

DMZ Europe 2016

About Ontologies –the Foundation of Data ModelingPeer M. Carlson, Dörffler & Partner

In a lot of texts and talks about Data Modeling,a reference is given to the concept of Ontolo-gies, whereas a further explanation of this noti-on is often missing. This presentation aims toreduce the knowledge gap of those who wantto learn more about the theoretical backgroundof Ontologies, to discover the interdisciplinaryrelationships in this area, and to understandhow modern Business Intelligence/Data Ware-house systems benefit from applying ontologicalideas.

The first part of the presentation briefly introdu-ces the Ontology from a philosophical point ofview – the study of being and existence in theworld and the interrelationships of objects. Inthe second part, the focus lies on Ontologies inthe field of Information Science as a formal spe-cification of a conceptualization – and fur-thermore, as a particular example, onOntologies in the context of the Semantic Web.Finally, the third part will show how Ontologiesserve as the foundation of Data Modeling – par-ticularly Data Vault – and how they facilitate theprocessing of unstructured data.

Peer M. Carlson is aBusiness Intelligence Con-sultant at Dörffler & PartnerGmbH, Germany. He hasbeen working on severalData Warehouse projects,mainly in the field of Micro-soft SQL Server as well as

Teradata. He has a lot of experience in Data Mode-ling, the Data Vault methodology, ETL process designand an in-depth knowledge of BI/DW architectures.Peer is a certified BI Expert (TDWI Europe) and a Cer-tified Data Vault 2.0 Practitioner, and he holds a un-iversity degree in Computer Science with focus onBusiness Information Systems.

9

DMZ Europe 2016

Business Information Modeling:A Methodology for Data-IntensiveProjects, Data Science andBig Data GovernanceTorsten Priebe, Simplity

The talk presents an integrated methodology tostructure and formalize business requirements inlarge data-intensive projects, e.g. data ware-house implementations, turning them into preci-se and unambiguous data definitions suitable tofacilitate harmonization and assignment of datagovernance responsibilities. The presented ap-proach places a business information model inthe center -- used end-to-end from analysis, de-sign, development, testing to data qualitychecks by data stewards. In addition, the talkshows that the approach is suitable beyond tra-ditional data warehouse environments, applyingit also to big data landscapes and data scienceinitiatives -- where business requirements analy-sis is often neglected. As proper tool supporthas turned out to be inevitable in many projects,software requirements and their implementationin the Accurity Glossary tool are also discussed.The approach is demonstrated based on examp-les from a number of large banking data ware-house projects.

Torsten Priebe has over 15years' experience in datamanagement and businessintelligence, of which hespent over 10 years in con-sulting. Since 2014, Torstenis CTO at Simplity, a dyna-mically developing interna-

tional consultancy focusing on data warehousing andbusiness intelligence. Torsten's responsibilities at Sim-plity include driving the requirements for the Accuritydata governance software suite and guiding Simpli-ty's consultants particularly in the areas of architec-ture and data modeling. Before Simplity, Torsten washeading the data management team at Teradata andthe business intelligence practice at Capgemini in Vi-enna respectively. Torsten is co-author of three books,published numerous articles and talks regularly at in-ternational conferences. In addition, Torsten is a regu-lar guest lecturer for business intelligence atUniversity of Regensburg, where he received his PhDwith a thesis on integrating structured and unstructu-red information with semantic technologies in 2005.

PS-3C Modeling:proposed way of agile EDW datamodeling on the Data LakeRogier Werschkull, DIKW

The last years we have seen the rise of the ‘DataLakes’. While they are technologically suitablefor the time variant / non volatile part of the‘traditional’ data warehousing as we know it,they do not automatically cover/fix the difficultpart of solid EDW data warehousing: building asubject-oriented and integrated layer on top ofthis raw data. So even in the ‘age of Data Lakes’data modeling is still needed and maybe evenmore than before: to bring Data-science relateddiscoveries into production we must have agood solution for the ‘data wrangling’ part thattakes up most of the time ...To solve this problem, I propose a new (layered)modeling approach for building this integratedEDW. This modeling approach consists of twolayers:

1) A persistent staging area (PS) in the Data la-ke: structured / semi-structured data is histo-rized, based in a (technical) source Primary /Unique Key. We must strive to build thisusing a columnar-oriented (and compressed)storage method. Then, using schema-on readfunctionality, we model the integrated EDWlayer:

2) A layer based on [Business Concepts], descri-bed by [Context] and linked together by[Connector-context] tables (3C). In is inspiredon Data Vault modeling with the main diffe-rences being shared during this talk.

Currently, I am in the process of proposing thismodeling approach in a major organization inthe Netherlands.

Rogier is a DWH / BI Con-sultant with 16 years of ex-perience in the field. Hestarted as a frontend de-veloper with Business ob-jects, moved to OLAP cubedevelopment and ETL de-velopment (Oracle ware-house builder, Kimball Datamarts /Data warehouse)

10

DMZ Europe 2016

and the last 5 years to DWH /ETL automation (usingData Vault, build an ETL generator at my previousemployer) and DWH architecture and data modeling.Recently he has built some experience using ‘Big Da-ta’ technology (Hadoop ecosystem) and in the DataScience related field. This combined experience hasgiven me a broad perspective in data architecture ingeneral and has given me ideas on how to improvethe current way of modeling the integrated EDW.

A Great Challenge:Designing A New DWH For A BankMurat Cetinkaya, ING Bank Turkey

We have finished our new IBM BDW based datawarehouse successfully in one and half years. Iwill discuss this project at ING Bank from a mo-deling perspective, including the cons and prosof industry models, best practices, source sys-tem challenges, modeling challenges, planning,and physicalization.

Murat Cetinkaya is workingat ING Bank Turkey as ITProduct Manager/DataWarehouse Architect. Mr.Cetinkaya is a data ware-house professional withbroad experience in big-si-zed data warehouse pro-

jects. He has detailed experience of designing andimplementing staging area, operational data storeand all aspects of the data warehouse developmentlifecycle; requirements gathering, both logical andphysical design of the data models, ETL design andprocessing. He has held modeling team leadershipand DWH architect positions in Is Bank (the biggestprivate bank of Turkey) and ING Bank Turkey. He is al-so an active member of the DAMA Turkey Chapter.

Data Modeling for NoTablesThomas Frisendal, TF Informatik

It is time for a change in the way we work withdata models. The database world is changingrapidly as NoSQL (not only SQL) and new physi-cal data models such as graphs, key/value, co-lumn stores and HADOOP make headway intodatabase development. But data modeling isstill seen as synonymous with SQL tables, nor-malization and entity / relationship diagrams.

Graphs (as in directed graphs) is the natural re-placement for the 40+ years old data modelingdiscipline with its bias on normalization and anabstract relational model, which never quitemade it to mainstream data modeling. The lega-cy is that of SQL-tables, primary and foreignkeys etc. as opposed to relations / “relvars”,which constitute the academic foundation forrelational database concepts.

We will cover subjects such as:

• Make data models for the new world of BigData and NoSQL

• Make data models that are visually compel-ling and highly communicative (also for non-technical audiences)

• Work visually with the functional dependen-cies (and the join dependencies) to controlredundancy and consistency.

• Feel good at not doing normalization!

If you are a data modeler, or a data scientist or adeveloper doing data modeling as you go, thenthis is for you.

Thomas Frisendal is an ex-perienced consultant withmore than 30 years on theIT vendor side and as anindependent consultant.Since 1995 he has primarilybeen working with datawarehouse projects. Hisarea of excellence lies within the art of turning datainto information and knowledge. Since 2005 he hasspecialized in business analysis, concept “harvesting”and mapping, modeling of information and data aswell as design of business intelligence solutions.Thomas Frisendal is a member of “The Data Ware-house Institute Denmark” (member of the board),“The International Association for Information andData Quality” (IAIDQ) and “The Cognitive Science So-ciety”. His approach to Information-driven BusinessAnalysis is “New Nordic” in the sense that it repres-ents the traditional Nordic values such as superiorquality, functionality, reliability and innovation bynew ways of conceptualizing the business. Thomas isan active writer and speaker, and also Chief DataWarehouse Architect at SimCorp. Thomas lives in Co-penhagen, Denmark.

11

DMZ Europe 2016

How to compose a fact-based modelinto any kind of schemaClifford Heath, Infinuendo

Fact-based modeling is not just a powerfulbusiness communication tool, it is the one truering of power for the data modeler. By workingwith a fully decomposed schema in “elementaryform”, we avoid making a commitment to anyparticular kind of composite schema. This pre-sentation will show how different compositionrules applied to a fact-based schema will gene-rate an entity-relational, star, snowflake, hierar-chical, object-oriented, XML or graph schema.

Clifford Heath is a compu-ter research scientist whohas long experience in thedesign and implementationof enterprise-scale softwareproducts, and in the use offact-based modeling. He isa Certified Data Manage-

ment Professional (CDMP) at Masters level, has pu-blished a number of papers in peer-reviewed scientificjournals, holds several patents, is the creator of theConstellation Query Language and is a participant inthe Fact Based Modeling Working Group. Clifford hasfrequently presented at chapter meetings of the DataManagement Association as well as at the NATO CAXForum and the European Space Agency, and is CTOand founder at Infinuendo.

Casetalk –Data Modeling by exampleMarco Wobben, Casetalk

Q: Data modeling is described as a craft and on-ce completed the results may even seem artful.Yet outsiders may see data modeling as abs-tract, time consuming or even unnecessary. Inmany cases the data modeler interviewsbusiness experts, studies piles of requirements,talks some more, and then, hocus pocus, pres-ents a diagram with boxes, crows feet, arrows,etc… Then the slow process begins to keep thediagrams up to date, explain what the diagramsbehold, and sometimes even data modelersthemselves may get lost while maintaining agrowing set of data models and requirements.

A: Fact based information modeling is the veryopposite of abstract. Fact based informationmodeling uses natural language which expres-ses facts that are intelligible for both businessand technical people. It does not require peopleto understand the modeler’s magical languageof boxes and arrows. Although models can bepresented in several diagramming notations,they can be validated in natural language at alltimes. This gives both data modelers, technicallyskilled people, and business people the benefitof having a well documented and grounded da-ta model. Therefore the method of Fact OrientedModeling, is also known as "Data Modeling byExample".

Presentation highlights• key elements of fact oriented modeling;• data modeling with facts;• visualizing the model;• validating and verbalizing;• transforming and generating output

(E.g.: SQL, Relational, UML, XSD, PowerDesi-gner, etc.).

Marco Wobben is directorof BCP Software and hasbeen developing softwarewell over 30 years. He hasdeveloped a wide range ofapplications from financialexpert software, softwareto remotely operatebridges, automating DWH generating and loading,and many back- and frontoffice and webapplications.For the past 10 years, he is product manager and leaddeveloper of CaseTalk, the CASE tool for fact basedinformation modeling, which is widely used in un-iversities in the Netherlands and across the globe.

From Conceptual to Physical DataVault Data ModelDirk Lerner, ITGAIN

The fictitious company PennyStockInc needs tobuild a new data warehouse for their tradingdepartment. Due to past experience (project fai-lures, overtime, etc.) the BI Competence Center(BICC) decided this time to build conceptual, lo-gical and physical data models. Conceptual mo-dels to gather all information about their trading

12

DMZ Europe 2016

business, logical models to collect their businessrequirements and cover both relational databa-ses and NoSQL solutions. For the physical datamodel the BICC of PennyStockInc decided tochoose data vault due to the agility, flexibilityand the ability to integrate both relational data-bases and NoSQL solutions.

Attendees of this session will be a fictitious partof the PennyStockInc BICC Team and will workon some exercises to build the new data ware-house.

This session explores• Basic principles of

• Conceptual modeling• Logical modeling• Physical modeling

• How to model from a conceptual via logicalto a physical data vault model

• How to build and access the physical datamodel on Relational-DB (EXASOL) and NoS-QL-DB (REDIS):• Design the data layout on REDIS. Identify

the keys to represent the objects and whichvalues this keys need to hold

• Design the relational data model• Access both the key-value store and the

relational database.

Dirk Lerner is a well experi-enced IT Consultant at IT-GAIN and is as team leadresponsible for the Compe-tence Center Data Architec-ture, Data Modeling andData Vault. For around 15years he has headed BIprojects and is considered an expert for BI architectu-res and data modeling. Dirk is an advocate of flexible,lean and easily extendable data warehouse principlesand practices. As a pioneer for Data Vault in Germanyhe published various publications, is an internationalspeaker at conferences and author of the bloghttp://www.datavaultmodeling.de.

Business Information ModelingFusing the fact-based approachClifford Heath, Infinuendo

Most modeling languages are effective for com-municating only with people who have beentrained to read and interpret the details. Howe-ver, fact-based languages are deeply rooted innatural speech. They build a complete vocabula-ry for the business – not just a glossary of terms– and can be used directly with untrainedbusiness personnel. This radically improves thedepth and breadth of the business dialogue.

As each new fact is encountered, the modeler

13

DMZ Europe 2016

adds appropriate elements to an Object RoleModel diagram, which makes it easy to verbali-ze to natural language. Alternatively, the newfact can be expressed directly in the structurednatural language form of the Constellation Que-ry Language, which is immediately readable tobusiness users but can also be understood bythe computer. This makes it possible to describeany factual situation precisely, to ask or answerany factual question, and to elaborate detailedbusiness rules about possible or allowable situa-tions.

This presentation will introduce you to the ap-proach and to these two languages, peeking atthe respective software tools which generatephysical schemas. A group exercise will challen-ge you to apply it to model a sample problem.The attendee will learn to use verbalization andfact-based analysis to understand any subjectmaterial, and to communicate it at a deep con-ceptual level with those less expert in the field.

Clifford Heath is a computerresearch scientist who haslong experience in the de-sign and implementation ofenterprise-scale softwareproducts, and in the use offact-based modeling. He isa Certified Data Manage-

ment Professional (CDMP) at Masters level, has pu-blished a number of papers in peer-reviewed scientificjournals, holds several patents, is the creator of theConstellation Query Language and is a participant inthe Fact Based Modeling Working Group. Clifford hasfrequently presented at chapter meetings of the DataManagement Association as well as at the NATO CAXForum and the European Space Agency, and is CTOand founder at Infinuendo.

Data Modeling forSustainable SystemsGraham Witt

Systems, like any other high-value acquisitions,should continue to work after the warranty peri-od, but far too often fail to work completely asintended. Data modelers can enhance the quali-ty and lifespan of a system if they take a broa-

der view than one which simply convertsinformation storage and retrieval requirementsinto a data model. This workshop looks at howto add value to the data modeler’s contributionto system development or package customizati-on, including:

• identifying real-world complexity and its im-plications for the choice of data structures

• the role of generalization in managing suchcomplexity

• development of a common vocabulary for re-al-world concepts, attributes, relationshipsand processes, and use of that vocabulary insystem artifacts and user interfaces

• recognizing how the data is to be used inBusiness Intelligence and the implications ofthat use for choice of data structures

• analysis of how changes in the real world areto be recorded in the chosen data structures,with implications for process and user inter-face design

• ensuring test plans cover data update side-effects adequately

• managing data model change effectively.

The workshop includes a series of case studiesdrawn from the speaker’s experience dealingwith these issues.

Graham has over 30 yearsof experience in deliveringeffective data solutions tothe government, trans-portation, finance and utili-ty sectors. He has specialistexpertise in business requi-rements, architectures, in-formation management, user interface design, datamodeling, database design, data quality and businessrules. He has spoken at conferences in Australia, theUS and Europe and delivered data modeling andbusiness rules training in Australia, Canada and theUS. He has written two textbooks published by Mor-gan Kaufmann: “Data Modeling Essentials” (withGraeme Simsion) and “Writing Effective Business Ru-les”, and writes monthly articles for the Business RuleCommunity (www.brcommunity.com).

14

DMZ Europe 2016

Advanced Data ModelingChallenges WorkshopSteve Hoberman,Steve Hoberman & Associates

After you are comfortable with data modelingterminology and have built a number of datamodels, often the way to continuously sharpenyour skills is to take on more challenging assi-gnments. Join us for a half day of tackling realworld data modeling scenarios. We will comple-te at least ten challenges covering these fourareas:• NoSQL data modeling• Agile and data modeling• Abstraction• Advanced relational and dimensional mode-

ling

Join us as in groups as we solve and discuss aset of model scenarios.

Steve Hoberman has trai-ned more than 10,000 peo-ple in data modeling since1992. Steve is known for hisentertaining and interactiveteaching style (watch outfor flying candy!), and orga-nizations around the globe

have brought Steve in to teach his Data ModelingMaster Class, which is recognized as the most com-prehensive data modeling course in the industry. Ste-ve is the author of nine books on data modeling,including the bestseller Data Modeling Made Simple.One of Steve’s frequent data modeling consulting as-signments is to review data models using his DataModel Scorecard® technique. He is the founder of theDesign Challenges group and recipient of the 2012Data Administration Management Association (DA-MA) International Professional Achievement Award.

Why Business Analysts and DataModelers should hold hands moreGeorge McGeachie, Metadata Matters

How integrated are your business analysis anddata modeling efforts? Share the experiences ofa lifetime of modeling, and improve the coordi-nation between your business analysts and data

modelers, enabling them to build a coherent, in-tegrated, set of requirements and models, redu-cing the probability of project failure.

Some of the questions that will be answered in-clude:

• Why does a business analyst need to under-stand or contribute to data models?

• Why does a data modeler need to under-stand or contribute to business analysis?

• Why designing XML messages is NOT datamodeling, it is process modeling.

• How do Business Rules influence data mo-dels?

• Where does all this fit with other types ofmetadata, including Enterprise Architecture?

• How does this impact your project manage-ment and governance?

• What implications does this have for yourmodeling tools?

George McGeachie hasspent his working life crea-ting, managing and linkingdata models, process mo-dels, and others. He encou-rages organizations toconnect and utilize theirmetadata islands, to reco-gnize the wealth of information contained in their da-ta models, to recognize that the creation of datamodels must form part of an integrated approach toimproving their business, and therefore recognize theimportance of avoiding the creation of islands of me-tadata in the first place.

Data Modeling forPolyglot PersistencePhill Radley, BT

This presentation will give an overview of howBritish Telecom is adapting its approach to datamodeling and data architecture to support pro-ject teams wanting to store their data in data-bases that are not based on a Relational DataModel. An important element being the use FactBased Models (specifically Object Role Modelingusing the NORMA tool) to develop conceptualmodels that can be used to join data stores suchas HDFS, Hive, Hbase and Orient DB.

15

DMZ Europe 2016

Agenda:• Polyglot persistence another name for big da-

ta, graph data, NoSQL....• Back to basics - Conceptual, Logical and

Physical Data Models• Basic Conceptual modeling using Object Role

Modeling• Hadoop (The Hive Metastore, File Formats

and Data Catalogs)• Schema on Read and Data Wrangling (using

Datameer)• Key Value Data Model (Time series Data in

Hbase)• Graph Data Model (OrientDB and Neo4J).

Phill Radley is a PhysicsGraduate with an MBA whohas worked in IT and com-munications industry for 30years , mostly with BritishTelecommunications plc. Heis currently BT’s Chief DataArchitect at their Martles-

ham Heath campus in the UK. Phill works in BT’s coreEnterprise Architecture team with responsibility forData Architecture across BT Group plc. He currentlyleads BT’s MDM and Big Data initiatives driving asso-ciated strategic architecture and investment road-maps for the business. His previous roles in BTinclude; 9 years as Chief Architect for PerformanceManagement solutions, ranging from UK consumerbroadband through to outsourced Fortune 500 net-works and hi-frequency trading networks. He hasbroad global experience including BT’s Concert globalventure in USA and 5 years as Asia Pacific BSS/OSS Ar-chitect based in Sydney.

The Journey fromER Modeler to Data ScientistAsoka Diggs, Intel Corp.

In this session, we will be talking about whatdata science is and isn’t, with particular focus onthe skills that a data scientist is expected tobring to bear on a problem. We will discuss theeducational and skill expectations of a datascientist and how those relate to the skills we’vedeveloped as ER Modelers. Going by job des-criptions, data science can easily read like “jackof all trades, and master of them all too”. Howdo we break that down into something more re-asonably achievable?

• What is data science and what is a datascientist?

• What do they do?• Do ER Modelers make good Data Scientists?• Are the skill sets complementary?• And more – bring your questions and let’s

get into them.

After 15+ years of experi-ence in a variety of datamanagement disciplines,including database admi-nistration, ER modeling, ETLdevelopment, and data ar-chitecture, I am transitio-ning into the world ofanalytics / data science. It turns out that there IS so-mething even more fun than ER modeling, and I’vefound it. Today, my primary interest is in the organi-zational transformation involved in adopting analyticsas a source of competitive advantage. How does anorganization get there? What needs to be done? Whatorganization design and leadership changes are nee-ded to get the most benefit from analytics? I nowspend my work life practicing these new analytic mo-deling skills, and teaching others how to participateand contribute to predictive analytic projects.

smartDATA and Industrie 4.0Dr. Siegmund Priglinger,dr.priglinger consulting

smartDATA for Industrie 4.0 is based onCAD/CAM, PLM and SE models, where differentmodeling methods/languages and tools arecombined and monitoring approaches for ex-changing and communication between the dif-ferent systems at the engineering and shop-floorlevels will be taken into account. Rigorous DataEngineering is a must. Formalization of schemamapping and matching is the key factor for In-dustrie 4.0.

More than 30 years experi-ence as product and soluti-on provider, companyfounder, IT Manager andConsulting Principal for ICTin different industries (Ma-nufacturing, Trade, Tele-communikation, IT, Energy,Healthcare, Public). Contact person for the manage-ment, the business divisions and the IT concerning“Business Intelligence” , “Data Governance” and “In-

16

DMZ Europe 2016

formation Management”. My current job is “to adviceother advisers” in this areas while I am project leaderof complex data migration or BI projects. I designednew methods to accelerate the migration of applicati-on models and to compare the quality of schema anddata of these application models. My vision is BI3:“Business Intelligence requires Best Information re-quires Best Integration.”

Building a Data Vaultusing an Analytical DatabaseJohan van der kooij, De Bijenkorf

Many data warehouses run on a ‘traditional’rowstore database, such as Oracle or SQL Server.With the rise of data driven organizations, thosedatabases might not be the best choice for BIand analytical purposes. That’s where dedicatedanalytical and column store databases come in-to play. How can data vault modeling be used insuch an analytical column store database? Chal-lenges such as end dating (using expensive up-date statements) and satellite splitting (mostlydone for storage and performance) suddenly canbe approached in a completely different way.This talk shows how HP Vertica runs at the Bi-jenkorf on an AWS infrastructure, utilizing thefull potential of the analytical database environ-ment combined with the power of data vaultmodeling.

Johan works as BI/Bigdataarchitect for the eCommercedepartment at the Bijen-korf, the largest premiumdepartment store in theNetherlands. The Bijenkorfis part of the Selfridgesgroup, an international lu-

xury retail with annual sales total over £2 billion. Wi-thin the eCommerce department, Johan works in theData Technology team. The team operates a datawarehouse environment which offers self- service re-porting, analytical capabilities and web integration tothe organization. Prior to the Bijenkorf, Johan workedfor several organizations in retail, finance and energyto help them become data driven.

Creating Ontologies fromyour Existing E/R Data ModelsNicolas Leveque, Deutsche Bank

During the last two generations, visionary peo-ple like Codd, Chen, Kent, Blaha, Wells, Inmon,and Hoberman contributed to the developmentand success of the Entity Relationship modeland its implementation by modern RDBMS ven-dors. During this period, thousands of students,scholars and professionals were taught the threedata modeling perspectives (Conceptual, Logicaland Physical), and data management has beco-me one of the most lucrative, job creating, andever changing IT industry.

The last couple of years, our industry has seenthe advent of a new modeling technique thatfacilitates the integration of heterogeneous datasources by resolving semantic heterogeneitybetween them: Ontology.

This presentation will explain in detail how toleverage all of the numerous E/R Data modelsproduced and turn them into an OWL ontologymodel. This session will explain the links bet-ween ontology and E/R model constructs, andwill help you with learning and creating yourfirst ontology from their existing E/R data mo-dels.

So, if, as with Deutsche Bank, your company hasstarted embracing Linked Data and ontologymodeling, come and learn how to reuse whatyou did and turn your E/R Data Models into abasic OWL ontology.

Nicolas Leveque is Head ofPhysical Data Architecture,Deutsche Bank, and a DataArchitect with 20 years ex-perience mainly within fi-nancial institutionsincluding UBS, Barclays,Societe Generale, CapitalGroup and Deutsche Bank. He specializes in EnterpriseData Architecture, Data Modeling and in bridging thegap between High Level architect and implementati-on teams. Still a hands-on technician, he doesn’t be-lieve in one-solution- fits-all, but strives to find theright solution depending on business and non-func-tional requirements. He also looks forward and sees

17

DMZ Europe 2016

how new practices and technologies could help tosolve data problems with a different and better mind-set.

Positioning Data Modelingin the Big Data AreaMartijn Imrich, Xomnia

So we arrived in the area of Big Data, and ever-yone is talking about Hadoop, Data Lakes andwhy we should have a data warehouse or dodata modeling anyway. And there’s no shortageof technology. Every day new techniques arrivewith fancy names such as Spark, Splunk, Hunk,Hive, Pig, YARN and so on. But how do we de-sign an effective Big Data Architecture? Andwhat is the role of Data Modeling in the Big Da-ta area?

A so called Reference Architecture has beencreated in many domains, ranging from applica-tions to Business Intelligence. Martijn will sug-gest a Reference Architecture for Big Data, andwhere data modeling fits.

Martijn Imrich is helping companies to be more DataDriven since the nineties. Today he is a Partner atXomnia, an Amsterdam based Big Data startup. Hismain activities are in innovation workshops, educati-

on and designing (Big) DataArchitectures. He’s the firstknown to publish a Big Da-ta Reference Architecture,positioning both Data Mo-dels and Technology. Beforejoining Xomnia, Martijnheld positions like LeadBusiness Analyst at Royal Philips to develop the Con-ceptual, Logical and Physical Data Models to supportthe new Philips application landscape in HealthTechservices. As Product Owner at Randstad, he introdu-ced Randstad to predictive analytics in Marketing andHR.

Modeling of Reference SchemesDr. Terry Halpin, Professor at INTIInternational University, Malaysia

This presentation illustrates the different ways inwhich a variety of data modeling approaches(ER, UML, ORM, RDB, OWL) support the import-ant task of identifying individuals of interest tothe business domain. In so doing, it also expo-ses the limitations and hurdles that might needto be overcome when transforming from oneapproach to the other. It addresses the followingcategories of reference scheme: (1) simple refe-rence schemes; (2) compound reference sche-mes involving multiple components; (3)

18

DMZ Europe 2016

disjunctive reference schemes involving optionalcomponents; and (4) context-dependent refe-rence schemes, where the preferred identifier foran entity varies according to the context.

In natural language, one usually refers to indivi-dual things by means of proper names (e.g. ‘Ba-rack Obama’) or definite descriptions (e.g. ‘Thecurrent president of the United States’). Langua-ges for data modeling and/or ontology modelingvary considerably in their support for such natu-ral reference schemes. An understanding of the-se differences is important both for modelingreference schemes within such languages andfor transforming models from one language toanother. This presentation provides a critical andcomparative review of reference scheme mode-ling within the Unified Modeling Language (ver-sion 2.5), the Barker dialect of EntityRelationship modeling, Object-Role Modeling(version 2), relational database modeling, andthe Web Ontology Language (version 2.0).Using illustrative examples, it identifies whichkinds of reference schemes can be captured wi-thin specific languages as well as those refe-rence schemes that cannot be. The analysis

covers simple reference schemes, compound re-ference schemes, disjunctive reference and con-text-dependent reference schemes.

Dr. Terry Halpin is a Rese-arch Fellow in ComputerScience at INTI Internatio-nal University (Malaysia),and a data modeling con-sultant. He previously heldsenior faculty positions incomputer science at TheUniversity of Queensland (Australia), and professor-ships in computer science at Neumont University(USA) and INTI International University. His prior in-dustrial experience includes many years in data mo-deling technology at Asymetrix Corporation,InfoModelers Inc., Visio Corporation, Microsoft Cor-poration and LogicBlox. His doctoral thesis formalizedObject-Role Modeling (ORM/NIAM), and his currentresearch focuses on conceptual modeling and rule-based technology. He has authored over 200 technicalpublications and nine books, and has co-edited ninebooks on information systems modeling research. Heis a member of IFIP WG 8.1 (Information Systems), isan associate editor or reviewer for several academicjournals, is a regular columnist for the Business RulesJournal, and is a recipient of the DAMA InternationalAchievement Award for Education and the IFIP Out-standing Service Award.

Be in the Zone!