Course description

The course starts at 09.30 am and ends at 5 pm. Registration commences at 08.30 am.

• What really is a data model?
• Essential components – entities, relationships, and attributes
• The basics of diagramming – Entity-Relationship Diagrams (“ERDs”)
• The narrative parts of a data model – definitions and assertions
• Group exercise – getting started on a data model, then refining it
• Common misconceptions about data models and data modeling
• The real purpose of a data model
• Three types of data models – different levels of details for different purposes
• Contextual, Conceptual, and Logical Data Models – purpose, audience, definition, and examples
• How data models help in process improvement, requirements definition, and reporting
• Forward- and reverse-engineering uses of data modeling
• Overview of a three-phase methodology for developing a data model
• References – books and useful web sites

• Top down vs. bottom up approaches to beginning a data model – when is each appropriate?
• Advantages of a bottom-up approach
• A bottom-up approach focusing on collecting and analyzing terminology
• A structure for sorting terms and discovering entities
• Exercise – developing an initial conceptual data model
• Entities – what they are and are not
• Guidelines for naming and defining entities
• Three questions to help you quickly develop clear, useful entity definitions
• Five criteria that entities must satisfy, and four common errors in identifying entities
• Exercise – identifying flawed entities
• Identifying relationships
• Fundamental vs. irrelevant or transitive relationships
• Good and bad relationship names
• Multiplicity or cardinality – 1:1, 1:M, and M:M relationships, and useful facts about each
• Common errors and special cases – recursive, multiple, and supertype-subtype relationships
• Attributes – guidelines and types
• Attributes in conceptual models vs. logical models

 

Phase 2 – Develop the initial logical data model by adding rigor, structure, and detail

• What’’s involved in developing a logical model – shifting the focus from entities to attributes
• Multi-valued, redundant, and constrained attributes, with simple patterns for dealing with each
• An understandable guide to normalization – first, second, and third normal forms
• Higher order (fourth and fifth) and Boyce-Codd normal forms
• Guidelines for a smooth progression from conceptual to logical
• Exercise – developing the initial logical data model
• Four types of entities – kernel, characteristic, associative, and reference
• Guidelines and patterns for dealing with each type of entity
• How to draw your E-R Diagram for maximum readability and correctness
• Optional and mandatory relationships
• Considering time and history when looking at relationships
• Six questions to ask whenver a data range appears in a data model
• Identifying and dealing with transitive relationships – clues and proof

• Attribute granularity – definitions of non-atomic and semantically overloaded attributes
• Guidelines for making non-atomic attributes atomic
• The perils of semantic overload, and what to do about it
• Dealing with derived attributes, and when to show them on the model
• A classword-based approach to attribute naming
• Typical attribute documentation
• A common source of confusion and disagreement – primary keys
• What primary keys are, what they’re really for, and three essential criteria
• Alternate and foreign keys
• Why meaningless primary keys are used, and guidelines for creating them
• Guidelines for reference data
• Pulling it together – key techniques and guidelines covered in the class so far
• Using event analysis to discover additional requirements
• Exercise – using event analysis and extending a data model
• Presentation by teams of their solutions
• How data modeling relates to process modeling, use cases, and services
• A layered framework for business analysts
• How other techniques (e.g., workflow modeling) support data modeling
• A three-step procedure for meeting new requirements
• Advice on extending the model in an orderly fashion
• Exercise – meeting new requirements on the data model
• Recap – contextual, conceptual, and logical data models
• Different skills and participants for conceptual vs. logical modeling
• How the modeler/analysts’s role changes as a project progresses
• A little philosophy for effective data modeling
• The four Ds of data modeling – definition, dependency, detail, and demonstration
• Wrap-up – the approach we followed throughout the class.