DAMA DAY 2016, September 15th (8:30-5:30) at Montgomery Park Auditorium
Contemporary Data Modeling with advanced topics
Presented by: Gordon Everest
We live in a world of Big Data, trying to make sense of it to deliver Business Intelligence. We have the promise of NOSQL and Applied Data Science to deliver some solutions. In dealing with today's challenges we need some new and different ways of thinking. Do we still need Conceptual models? How do they continue to benefit us? What are the limitations of relational models, and how do we overcome them? What are the best strategies for preparing and presenting data model diagrams to support meaningful conversations with our business partners? How do our data modeling tools help? How is the data warehouse environment evolving? Are we beyond dimensional models? How do we work with NOSQL tools and schema-less databases? Do these environments still need data modeling? What is a Data Lake and how do we manage it?
Dr. Everest will explore these questions and provide some insight for dealing with issues surrounding the role of data modeling in the world we live in today. His presentation will provide some new strategies and approaches. His presentations are always lively and will challenge our thinking.
DATA MODELING – beyond the basics, and beyond relational
- Data Modeling Schemes – semantics vs. syntax; constructs, rules of composition, and constraints
- Types of models, a taxonomy – the many faces of databases
- Where does the Relational Data Model fit?
- Conceptual vs. Logical vs. Physical data models – definitions, how do we differentiate?
- When does a conceptual model become a logical model?
- When does a logical data model turn into a physical model? Where is the dividing line?
- What role do conceptual models play? Do we still need them?
ADVANCED DATA MODELING CONCEPTS AND CONSTRUCTS
- Problems with the Relational data model:
- "TABLE THINK -- many of us (experienced data modelers) suffer from "Tableitis"
- Limited to at most 1:Many binary relationships (foreign keys)
- Representing relationships in two different ways
- Need for normalization
- Focusing on populations of "things" and relationships among those things
- So where do Attributes fit in?
- Representing relationship in one consistent way; handling M:N relationships directly
- Ternary++ relationships, representing exclusivity and dependency constraints unambiguously
- Nine criteria for picking identifiers, surrogate keys
- Introduction to Subtypes and Supertypes, and constraints on them
NORMALIZATION – Myths. The biggest problem with ER/Relational/Object modeling.
- Proper interpretation and application (in a relational environment)
- Could you explain the rules of normalization (first 3 rules?, 5?, 6?) to a novice data modeler?
- "New" and simpler methods to correctly and completely normalize a data structure
- storing attributes with their determinant(s)
- Exercises to test your understanding
DATA MODEL PRESENTATION (getting started)
- Guidelines and strategies for presenting data model diagrams to business users
- Differentiation and layout – the good and the bad
- Horizontal and vertical abstraction
- Unfolding levels of detail with some examples
DATA MODEL PRESENTATION (continued)
- Conceptual models replaced with navigation and abstraction tools based on an underlying, fully detailed, (logical?) business data model.
FOCUS ON DATA WAREHOUSING and more
- Historical perspective: Bill Inmon vs. Ralph Kimball, how do they differ?
- Architecture of the Operational environment – ODS, staging (ETL), where is data cleansing?
- Dimensional modeling, fact and dimension tables, cube, star schema and snowflake
- Fact tables and dimension tables - should they be normalized?
- Business Intelligence (BI), Analytics, data mining, OLAP
- And now the "Data Lake" (from Bill Inmon) or perhaps a "Data Swamp"!
- NOSQL – rationale and motivation, Schema-less or "schema on read"?
- Type of Tools: Key value pair, multi-column, (directed?) graph, "document" (XML…)
- Data modeling in a NOSQL environment. Do we still need data modelers?
- Emerging role of the data analyst/scientist
Dr. Everest is actively engaged in the world of data modeling and data management, speaking to computer professional societies and consulting with large and small organizations. Early retirement has allowed him to focus on helping organizations improve their understanding and approach to data modeling. He continues to teach a course in advanced data modeling at the University of Minnesota. It is fully online and available to working professionals as well as students. (http://geverest.umn.edu).
He is an astute observer and interpreter of developments and trends in information technology, data management, and particularly business data modeling.
For more information go to: http://geverest.umn.edu/home/brief-biography
|September 15, 2016 (Thursday)
8:30am - 5:30pm
Montgomery Park Ballroom
2791 NW Vaughn St., Portland, OR 97210
Early Bird Registration Is Over
Regular Registration (Starting Sept 12th)
$129 DAMA PDX Members & Employees of
DAMA PDX Corporate Members
Upcoming DAMA PDX Metro Events