Skip to main content
Generic filters
Search in title
Search in content
Search in excerpt

Data Model


A data model is an abstract framework or blueprint for organizing and defining how data is collected, stored, processed, and presented. It conceptually represents the relationships among different data elements and how they can be processed to provide valuable insights.

Data modeling is critical to many IT initiatives, as it can help ensure that data is collected, stored, and processed to meet the organization’s needs. It can also help improve data quality, reduce redundancy, and make it easier to integrate systems and share data across them. 

Solid “data plans” for collecting and storing data are critical for all systems development work. Once data is organized a certain way, any changes to the data setup would cause lots of rework. Data modeling helps ensure that business data is managed efficiently for the organization and its needs. 

Designated data specialists can do data modeling, but Business Analysts can also play a key role in data modeling. We included the video clip below because it covers data modeling from the BA perspective.

A Brief History

Data modeling has its roots in the 1960s, with the advent of the first database management systems. The earliest data models were hierarchical and network models. However, the relational model, proposed by Edgar F. Codd in 1970, revolutionized how we interact with data.

Codd’s model used tables to organize data elements and establish relationships between them, forming the foundation of relational database systems like Oracle, MySQL, and PostgreSQL.

The relational model made data modeling more accessible and applicable to various business contexts. 

Data Modeling Components

There are several key components and considerations in a data model, including:

  1. Entities: These are the subjects or objects in the system that data is being collected on. They could be physical entities (like a customer or a product) or conceptual entities (like an event or a transaction).
  2. Attributes: These are the individual pieces of data that are collected on each entity. For example, for a customer, the attributes might include their name, address, email, etc.
  3. Relationships: These describe how different entities interact with each other. For example, a customer (entity) might place an order (relationship) for a product (another entity).

Data Models and Their Abstraction Levels

Data models are used at different levels of data abstraction, providing a different view of data based on the audience’s needs:

  1. Conceptual data model: This defines what the system contains, providing a high-level view of data and its relationships. Business stakeholders and data architects often use it to describe the general structure of the data without getting into specific details.
  2. Logical data model: This defines how the system should be implemented, regardless of the database management system (DBMS). It adds more details to the conceptual model, including attributes and primary and foreign keys.
  3. Physical data model: This defines how the system will be implemented using a specific DBMS, programming language, or data storage technology. It includes all necessary physical storage considerations, constraints, and indexes needed.

Different types of data models are often used in different stages of a system’s development and implementation.

For example, a conceptual data model is often used during the planning phase, a logical data model during the design phase, and a physical data model during the implementation phase. 

Types of Data Models: From Hierarchical to Graph

Over the years, data models have evolved to address different needs and work with different data types.

Beyond the relational model are object-oriented models, which use objects and classes to represent data, and hierarchical models, which organize data in a tree-like structure. Network data models extend hierarchical models by allowing child nodes to have multiple parent nodes. 

More recently, NoSQL databases have brought new data models, including document-based models, key-value pairs, wide-column stores, and graph databases.

Graph data models, in particular, are gaining popularity due to their ability to represent complex relationships more naturally and efficiently, such as social network connections or web page links.

Data Modeling in the Era of Big Data and AI

Data modeling is more critical than ever in today’s big data and artificial intelligence age. With exponentially increasing volume, velocity, and variety of data, well-designed data models can help manage this complexity and unlock valuable insights.

For instance, they can help structure unstructured data, which makes up a significant portion of big data, making it usable for machine learning algorithms. 

At the same time, AI technologies, like automated machine learning, are now being used to automate data modeling and database design, helping organizations manage their data more effectively and make faster, data-driven decisions.

Data Models and Data Governance

Data modeling is also critical to data governance – the overall management of data availability, relevancy, usability, integrity, and security in an organization. A well-defined data model can provide a consistent framework for ensuring data quality and compliance, managing data assets, and establishing responsible parties for different data segments. 

Having a solid data model in place can also make it easier to audit data, identify and address security risks, and ensure that the organization complies with applicable laws and regulations regarding data use and privacy. 

The Future of Data Modeling

As we move into an increasingly digital future, the role of data modeling will likely continue to grow. With emerging technologies like IoT, blockchain, and quantum computing set to generate new data types at unprecedented scales, the demand for sophisticated, adaptable data models will only increase.

Data modeling will continue to evolve to meet these challenges and help organizations navigate their digital transformation journeys. 

Data Modeling for the Modern Data Stack – 10 mins  

YouTube player

How to get started with Data Modeling – 40 mins  

YouTube player