Data Modeling Fundamentals

Data Models

• Data modeling is an iterative process for creating a specific data model for a particular domain.
• Data models are simple representations (typically graphical) of complex real-world data structures.
• Data models show how data elements connect and how they are processed and stored.
• Models are abstractions of real-world objects or events.

Learning Outcomes

• Data modeling and why data models are important
• Basic data-modeling building blocks
• Business rules and their influence on database design
• Evolution of major data models
• Classification of data models by their level of abstraction

Introduction

• Designers, programmers, and end-users see data differently.
• Different perspectives on data can lead to designs that don't reflect organizational operations.
• Data modeling simplifies database design.
• Varying degrees of data abstraction reconcile different views of the same data.

Importance of Data Models

• Data models act as communication tools between designers, programmers, and end-users.
• They provide a comprehensive view of the database.
• Data models organize data for various users.
• Data models are abstractions that support the creation of well-designed databases.

Data Model Basic Building Blocks

• Entity: A unique and distinct object used for data collection and storage (e.g., a person, place, thing, or event).
• Attribute: A characteristic of an entity (e.g., name, address).
• Relationship: Describes the association between entities (e.g., one-to-many, many-to-many, one-to-one).
• Constraint: A restriction placed on data (e.g., salary range, GPA limits).

Example

• Entity: CUSTOMER, RECEIPT, MUSICAL CONCERT, ROOM
• Attribute: CUSTOMER NAME, CUSTOMER PHONE NUMBER, CUSTOMER ADDRESS
• Relationship: An agent can serve many customers; a customer can be served by one agent
• Constraint: Employee salary must be between 6,000 and 350,000; student GPA between 0.00 and 4.00; each class has one teacher.

Business Rules

• Descriptions of policies, procedures, or principles within a specific organization.
• Descriptions of operations to create/enforce actions within an organization's environment.
• Descriptions of data characteristics as viewed by the organization.
• Business rules must be written, kept up to date, easy to understand and widely disseminated.

Example (Business Rules)

• A customer may generate many invoices.
• Each invoice is generated by only one customer.
• A training session cannot be scheduled for fewer than 10 or more than 30 employees.

Sources of Business Rules

• Company managers
• Policy makers
• Department managers
• Written documentation
• Direct interviews with end-users

Reasons for Identifying and Documenting Business Rules

• Standardize the company's view of data
• Act as a communication tool between users and designers
• Help designers understand data nature, role, scope, and business processes.
• Support development of appropriate relationship participation rules and constraints.
• Facilitate creation of accurate data models

Translating Business Rules into Data Model Components

• Nouns translate to entities.
• Verbs translate to relationships between entities.
• Relationships are bidirectional.
• Questions to determine relationship types: How many instances of B are related to one instance of A? How many instances of A related to one instance of B? Example: How many classes can one student enroll in? How many students can enroll in one class?

Naming Conventions

• Names should be unique and distinguishable from other objects.
• Names should be descriptive and familiar to users.
• Proper naming facilitates communication and promotes self-documentation.

Data Modeling and Data Models (continued)

• Data modeling involves creating specific data models for a determined domain.
• Data models are usually graphical representations of the complex real-world structures.

The Evolution of Major Data Models

• First Generation (1960s-1970s): File system.
• Second Generation (1970s): Hierarchical.
• Third Generation (Mid-1970s): Network.
• Fourth Generation (Mid-1980s): Relational.
• Fifth Generation (Mid-1990s): Object-oriented, Object/relational, XML, Hybrid DBMS.
• Emerging Models (Late 2000s to Present): Key-value store, Column store, NoSQL.

Hierarchical Models

• Manage large amounts of data for complex projects.
• Data is represented by an upside-down tree.
• Depicts one-to-many relationships.

Network Models

• Represent complex data relationships effectively.
• Improve database performance and establishes standards.
• Depicts many-to-many and one-to-many relationships.

Data Model Basic Building Blocks (continued)

• Constraints: Specific rules that restrict data (e.g., salary ranges, date formats, data types)

Standard Database Concepts

• Schema: Conceptual organization of the entire database.
• Subschema: A portion of the database viewable by specific application programs.
• Data Manipulation Language (DML): Environment for data management.
• Schema Data Definition Language (DDL): Language enabling schema definition.

The Relational Model

• Developed by E.F. Codd at IBM in 1970.
• Replaced standard transmission databases.
• Tables are matrices of rows and columns (relations).
• Rows are called tuples.

Relational Database Management System (RDBMS)

• Performs functions of hierarchical and network DBMSs.
• Simplifies understanding and implementation of relational data models.
• Hides relational model complexities from the user.
• Relational diagram illustrates entities, attributes, and relationships.
• Relational tables store related entities.

The Entity Relationship Model (ERM)

• Widely accepted standard for data modeling.
• Introduced by Chen in 1976.
• Graphical representation of entities and their relationships in a database structure.
• Graphic representations of database components.
• Entity instance represents an occurrence or row in a relational table.
• Entity set is a group of similar entities.
• Connector words classify relationship types.

Object-Oriented Data Model (OODM)

• Combines data and relationships in a single structure (object).
• OODM is the basis for OODBMS.
• Objects have attributes and operations.
• Objects are self-contained building blocks.
• Objects are abstractions of real-world entities.

Object-Oriented Data Model (OODM) (continued)

• Attributes describe object properties
• Objects with shared characteristics are grouped into classes.
• Classes form a hierarchy via inheritance (methods and attributes from parent classes).
• UML (Unified Modeling Language) graphically models objects, classes and other components.

Object/Relational and XML

• Extended Relational Data Model (ERDM): Supports object-oriented features and complex data representation.
• Extensible Markup Language (XML): Manages unstructured data for efficient exchange.

Big Data, Challenges, and Technologies

• Big Data aims to manage large volumes of web and sensor-generated data.

• Volume, velocity, and variety are key characteristics of big data.

• Big Data challenges include managing the volume, speed, and variety of data, as well as being expensive and inconsistent.

• Hadoop Distributed File System (HDFS) and MapReduce are new technologies designed to handle big data.

• NoSQL databases are used for high availability, fault tolerance, and storing a large number of sparse data.

NoSQL Databases

• Amazon, Facebook, YouTube, and Google use NoSQL.
• Relational model is not used in NoSQL
• NoSQL is based on distributed architectures.
• It supports scalability, availability, and fault tolerance.
• High performance is favored over transaction consistency.
• Data is commonly stored using a key-value store.

The Physical Model

• Represents data at its lowest level of abstraction.

• Details the actual storage and access to data on storage media (disk, tape).

• Relational models operate at the logical level and do not require physical details.

Data Abstraction Levels

• External Model: End-user views.
• Conceptual Model: Organization's global view.
• Internal Model: Database's specific representation for a DBMS.
• Physical Model: Lowest level of abstraction, how data is stored on disks or tapes.

Data Modeling Components

• Degree of Abstraction: The level of detail, e.g. low or high.
• Focus: Data aspect of the model, e.g, internal aspects like the storage mechanism vs. more user-oriented attributes.
• Independent of: Hardware, software, users, etc

Review Questions / Summary

• Questions were presented in slides about important elements like "Why are data models important?", "What is database structure?", "What are business rules?", "What are the sources of business rules?", and more.
• Summaries on slides give overviews of different models, data structures and data abstraction levels, concepts and principles related to these models and provide key components of database design like entities, attributes, relationships, business rules and constraints.

Data Model Basic Terminology Comparison

• Presents various data model types (e.g., File, Hierarchical, Network, Relational, ER, OO)
• Compares elements of those models for each (e.g., example data type, name field, record type).

This quiz explores the fundamental concepts of data modeling, including key building blocks, business rules, and their impact on database design. It also covers the evolution of data models and their importance in handling Big Data. Test your knowledge on the various data abstraction levels and industry usage by major tech companies.