3. Data Models

Data Modeling

Process of creating a data model for a specific problem; iterative and progressive.

Data Model

Simplified representation of complex real-world data structures.

Model

Abstraction of a complex real-world object or event.

Implementation-Ready Data Model

Includes data structure, enforceable integrity rules, and data manipulation methods.

Importance of Data Models

Improve communication, requirements gathering, data organization, and database design.

Four Basic Building Blocks

Entities, attributes, relationships, and constraints.

Entity

Person, place, thing, idea, or event about which data is stored.

Relationship

Association between entities.

Constraint

Restriction placed on data to ensure data integrity.

One-to-Many (1:M)

One entity instance relates to many instances of another entity.

Many-to-Many (M:N)

Multiple instances of each entity can relate to multiple instances of the other.

One-to-One (1:1)

One entity instance relates to one instance of another entity.

Business Rule

Brief, precise, unambiguous description of a business policy, procedure, or principle.

Business Rule Language

Business-focused and nontechnical.

Business Rule Sources

Executives, managers, policymakers, documentation, and end-user interviews.

Importance of Business Rules

Standardize data views and help identify entities, relationships, and constraints.

Nouns in Business Rules

Generally represent entities.

Verbs in Business Rules

Generally represent relationships.

Bidirectional Relationships

Relationships must be evaluated in both directions between entities.

Entity Naming Conventions

Use descriptive, familiar, singular nouns.

Attribute Naming Conventions

Use clear, descriptive names such as CUSTOMER_NAME.

Atomic Attributes

Attributes broken into individual component parts.

Unique Identifier

Value that uniquely distinguishes one record from all others.

Hierarchical Model

Early data model organized as an upside-down tree.

Hierarchical Model Disadvantages

Complex implementation, difficult management, structural dependence, and lack of standards.

Network Model

Similar to hierarchical but allows records to have multiple parents.

Network Model Disadvantages

Cumbersome, structurally dependent, and limited ad hoc querying.

Schema

Description of the database structure.

Subschema

Description of a specific view of the database.

DDL

Data Definition Language; defines database structures.

DML

Data Manipulation Language; manipulates stored data.

Relational Model

Data model based on set theory that organizes data into tables.

Relation

Table consisting of rows and columns.

Tuple

One row in a relational table.

Attribute in Relational Model

One column in a relational table.

RDBMS (Relational Database Management System)

Manages data stored in related tables.

SQL (Structured Query Language)

Used to query and retrieve relational data.

SQL Engine

Executes database queries.

Relational Schema/Diagram

Representation of relations, attributes, and relationships.

Entity Relationship Diagram (ERD)

Graphical representation of database entities and relationships.

Entity Instance/Occurrence

One individual record or row representing an entity.

Connectivity

Relationship classification: 1:1, 1:M, or M:N.

Crow's Foot Notation

Commonly used notation for entity-relationship diagrams.

UML (Unified Modeling Language)

Commonly used for object-oriented models.

Object-Oriented Model

Stores data and relationships together in self-contained objects.

Object

Self-contained structure representing a real-world entity.

Semantic Data Modeling

Modeling approach intended to represent the real world more closely.

Extended Relational Data Model (ERDM)

Combines relational modeling with object-oriented features such as inheritance.

XML (Extensible Markup Language)

Supports data exchange and varied data structures.

Internet of Things (IoT)

Network of Internet-connected devices that collect and exchange data.

Big Data

Managing rapidly growing, varied data to generate business insights.

Big Data 3Vs

Volume, Velocity, and Variety.

Volume

Large amount of data being stored.

Velocity

Speed at which data grows and must be processed.

Variety

Multiple formats and types of source data.

Big Data Challenges

High storage and processing costs; difficulty handling large or unstructured data.

Key-Value Database

Stores key-value pairs for fast retrieval, such as shopping cart sessions.

Document Database

Stores JSON-like documents, such as customer profiles.

Graph Database

Models relationships, such as social networks and recommendations.

Cloud Storage

Scalable storage using cloud services.

Distributed Processing

Processes large datasets across multiple computers.

Data Lake/Lakehouse

Combines flexible data storage with analytics and management.

AI Analytics

Uses AI and machine learning to identify patterns and predict outcomes.

Data Abstraction

Representing data at different levels of detail.

ANSI/SPARC Framework

Defines external, conceptual, and internal levels of data abstraction.

External Model

End-user view of data, usually focused on a specific business unit.

External Schema

Specific representation of an external view.

Conceptual Model

Global, high-level view of the entire organization's database.

Conceptual Schema

Graphical representation of the conceptual model.

Logical Design

Task of creating a conceptual model.

Software Independence

Model does not depend on specific DBMS software.

Hardware Independence

Model does not depend on specific hardware.

Internal Model

Database structure as viewed by the DBMS.

Internal Schema

Database structure expressed using a specific DBMS's constructs, such as SQL.

Logical Independence

Internal model changes do not affect the conceptual model.

Physical Model

Lowest abstraction level; describes how data is physically stored.

Physical Independence

Physical storage changes do not affect the internal model.

Levels of Data Abstraction

External → Conceptual → Internal → Physical.

Good Data Model

Conceptually simple without sacrificing database completeness.