What is Data Modelling? Explained Briefly

Data is everywhere. We use it to make decisions, communicate, learn, and create. But how do we store and manage data in a way that makes sense and meets our needs? This is where Data Modelling comes in. Data Modelling is the art and science of designing and organising data structures to support various applications and systems. In this blog, we will answer the question: What is Data Modelling? We will also explore the history, types, processes, advantages, challenges, and Data Modelling examples.  

Table of Contents 

1) Understanding What is Data Modelling 

2) History of Data Modeling 

3) Different types of Data Models 

4) Process of Data Modeling 

5) Advantages of Data Modeling 

6) Challenges of Data Modeling 

7) Examples of Data Modeling 

8) Conclusion 

Understanding What is Data Modelling  

Data Modelling is creating a representation of data structures, usually diagrams, schemas, or models. Data Models describe the properties, relationships, and constraints of Data elements and the operations and rules that can be applied to them. Data Models can be used for various purposes, such as:  

a) Designing Databases: Data Models help define the structure, format, and integrity of data stored in Databases. They also help optimise Databases' performance, security, and scalability.  

b) Developing applications: Data Models help specify applications' data requirements, logic, and functionality. They also help to ensure compatibility and interoperability between different applications and systems.  

c) Analysing data: Data Models help to understand the meaning, context, and quality of data. They also help to identify patterns, trends, and insights from data.
 

History of Data Modelling  

Data Modelling has a long and rich history, dating back to the 1950s. Some of the major milestones in the evolution of Data Modelling are:  

a) 1950s: The first Data Models were based on the network and hierarchical models, representing data as interconnected nodes or nested trees. These models, such as IBM's IMS, were mainly used for navigational Databases.  

b) 1960s: The Entity Relationship (ER) model, which represented data as entities, attributes, and relationships, was introduced by Peter Chen. This model was mainly used for conceptual and logical Data Modelling and Relational Databases.  

c) 1970s: The relational model, which represented data as tables, columns, and rows, was proposed by Edgar Codd. This model was mainly used for physical Data Modelling and relational Databases, such as Oracle and SQL Server.  

d) 1980s: The object-oriented model, which represented data as objects, classes, and methods, was developed by Michael Stonebraker. This model was mainly used for object-oriented Databases like ObjectStore and O2.  

e) 1990s: The object-relational model was created by combining the features of the object-oriented and relational models. This model was mainly used for object-relational Databases like PostgreSQL and Oracle.  

f) 2000s: The XML model, which represented data as elements, attributes, and text, was adopted by the World Wide Web Consortium (W3C). This model was mainly used for XML Databases like eXist and MarkLogic.  

g) The 2010s: The NoSQL model was popularised by the rise of big data and cloud computing, representing data as key-value pairs, documents, graphs, or columns. This model was mainly used for NoSQL Databases, such as MongoDB, Neo4j, and Cassandra.  

Dive into Database mastery – Join our Database Training today! 

Different types of Data Models  

Data Models can be classified into three main types, depending on the level of abstraction and detail they provide:
 

Conceptual model 

A conceptual model is a high-level and simplified representation of data that focuses on a domain's main concepts, entities, and relationships. It is independent of any specific technology, application, or system. It is used to communicate a project's data requirements and scope to stakeholders, such as business analysts, domain experts, and users.  

Logical model 

 A logical model is a detailed and structured representation of data, which focuses on the properties, attributes, and constraints of data elements, as well as the operations and rules that can be applied to them. It is independent of any specific technology but dependent on the Data Model type, such as relational, object-oriented, or NoSQL. It is used to design a system or application's data structures and logic, such as Databases, APIs, or interfaces.  

Physical model 

 A physical model is a concrete and optimised representation of data, which focuses on the implementation, format, and storage of data elements, as well as the performance, security, and scalability of the system or application. It depends on the specific technology, platform, or tool, such as SQL, MongoDB, or Neo4j. It is used to create and maintain a system or application's physical data structures and components, such as tables, indexes, or clusters. 

Process of Data Modelling  

Data Modelling is an iterative and collaborative process which involves the following steps:
 

Recognise the entities 

An entity is a person, place, thing, or concept that can be identified and distinguished from others. For example, in a library system, some entities are book, author, publisher, and borrower. The first step of Data Modelling is identifying and listing the entities relevant to the domain or problem.  

Determine key properties for each entity  

A property is a characteristic or feature that describes an entity. For example, a book entity's properties are title, genre, ISBN, and price. The second step of Data Modelling is to determine and list the essential and unique properties of each entity.  

Establish relationships among entities 

A relationship represents a link or association among two or more entities. For example, a book entity has a relationship with an author entity, expressed as “an author writes a book”. The third step of Data Modelling is to establish and list the relationships among the entities, as well as the cardinality and optionality of each relationship. The cardinality is the number of instances that can participate in a relationship, such as one-to-one, one-to-many, or many-to-many. Optionality is the degree of participation or dependency of an entity in a relationship, such as mandatory or optional.  

Fully map attributes to entities 

An attribute is a specific value or instance of a property that belongs to an entity. For example, some of the attributes of a book entity are “Harry Potter and the Philosopher’s Stone”, “Fantasy”, “9780747532699”, and “£5.99”. The fourth step of Data Modelling is to fully map and list the attributes that correspond to each property and entity and the data type and format of each attribute. The data type is the category or classification of data, such as string, integer, or date. The format is the representation or appearance of data, such as alphanumeric, decimal, or currency.  

Allocate keys 

A key is a property or a combination of properties uniquely identifying an entity or a relationship. For example, a book entity can be identified by its ISBN property, which is a unique identifier. A key can be either primary or foreign. A primary key uniquely distinguishes an entity in a table or collection, while a foreign key refers to the primary key of a different entity or relationship. The fifth step of Data Modelling is to allocate and list the keys needed to ensure the uniqueness and integrity of data and the referential integrity of relationships.  

Conclude and validate the Data Model 

The final step of Data Modelling is to conclude and validate the Data Model by checking for completeness, consistency, and accuracy. Completeness means that the Data Model covers all the data requirements and scope of the project. Consistency means that the Data Model follows the same standards, conventions, and rules. Accuracy means that the Data Model reflects the reality and logic of the domain or problem. The Data Model can be validated using various methods, such as reviews, feedback, testing, or simulation.  

Advantages of Data Modelling  

Data Modelling has many advantages, such as:  

a) Improves data quality: Data Modelling helps to ensure that data is accurate, complete, consistent, and reliable by defining the properties, constraints, and rules of data elements and relationships.  

b) Enhances data efficiency: Data Modelling helps to optimise the performance, security, and scalability of data by designing the structure, format, and storage of data elements and components.  

c) Facilitates data integration: Data Modelling helps enable the compatibility and interoperability of data by specifying the data requirements, logic, and functionality of systems and applications.  

d) Supports data analysis: Data Modelling helps to understand the meaning, context, and quality of data by describing the concepts, entities, and relationships of a domain or problem.  

e) Simplifies data communication: Data Modelling helps to communicate a project's data requirements and scope by using diagrams, schemas, or models that are easy to read and understand.  

Challenges of Data Modelling  

Data Modelling also has some challenges, such as:  

a) Complexity: Data Modelling can be complex and time-consuming, as it involves many steps, decisions, and details. It also requires knowledge, skills, and experience in Data Modelling techniques, tools, and languages.  

b) Variability: Data Modelling can be affected by the changes and variations in the data requirements, scope, and sources of a project, as well as the Data Model type, technology, and platform used. It also requires constant updates and revisions to keep up with the changes and ensure data validity and consistency.  

c) Diversity: Data Modelling can be difficult to standardise and harmonise, as it involves different Stakeholders, perspectives, and objectives. It also requires collaboration and communication among the Data Modellers, Developers, Analysts, and users and the alignment and integration of different Data Models, systems, and applications.
 

Start your data journey – Join our Introduction to Database Training today! 

Examples of Data Modelling  

To illustrate the concepts and principles of Data Modelling, here are some examples of Data Models based on different Data Model types:  

Network model 

 A network model represents data as interconnected nodes with multiple parent and child nodes. Each node consists of a record type, which defines the properties and relationships of the node, and a set of occurrences, which are the instances of the record type. A network model suits complex and hierarchical data, such as organisational structures, product catalogues, or genealogy trees. Here is an example of a network model for a library system: 
 

Hierarchical model 

A hierarchical model represents data as nested trees with one root node and multiple child nodes. Each node consists of a record type, which defines the properties and relationships of the node, and a set of occurrences, which are the instances of the record type. A hierarchical model is suitable for simple and hierarchical data, such as file systems, menus, or XML documents. Here is an example of a hierarchical model for a library system: 
 

ER (Entity Relationship) model 

 An ER model represents data as entities, attributes, and relationships. An entity is a person, place, thing, or concept that can be identified and distinguished from others. An attribute is a characteristic or feature that describes an entity. A relationship represents a link or association among two or more entities. An ER model is suitable for conceptual and logical Data Modelling, as well as for relational Databases. Here is an example of an ER model for a library system:
 

Relational model 

A relational model represents data as tables, columns, and rows. A table is a collection of data elements related to each other. A column is a property or attribute describing a table's data elements. A row is a specific value or instance of the data elements in a table. A relational model is suitable for physical Data Modelling, as well as for relational Databases. Here is an example of a relational model for a library system:
 

Object-oriented Database model 

An object-oriented Database model represents data as objects, classes, and methods. An object is an instance of a class, which is a collection of data and behaviour that defines the properties and functionality of the object. A method is a function or procedure that performs an action or operation on the object. An object-oriented Database model suits complex and heterogeneous data, such as multimedia, documents, or spatial data. Here is an example of an object-oriented Database model for a library system:
 

Unlock the power of data: Join our Relational Databases & Data Modelling Training today!

Conclusion  

In this blog, we have answered the question: What Is Data Modelling? We have also explored the history, types, processes, advantages, challenges, and examples of Data Modelling. Data Modelling is a vital and valuable process that helps design and organise data structures to support various applications and systems.