• Home
  • Introduction to Data Modeling

Introduction to Data Modeling

  • Rafat
  • March 15, 2022

In this article, we will understand a little more about data modeling with respect to learning Power BI.

What Is Data Modeling?

Data modeling is the act of using words and symbols to represent data and how it flows in a simplified representation of a software system and the data bits it includes. Data models serve as a roadmap for creating a new database or altering an existing one. It is a process of defining how your tables are related to each other. It is also where you create your calculated columns, calculated tables, custom measures, etc. Data modeling enables a firm to make the most use of its data in order to satisfy business information demands. It enables business and technical workers to collaborate on how one can store, access, share, update and use data inside of an organization.

What Do We Need Data Modeling For?

Why is data modeling done in the first place? Few of its objectives are the following:

  • Data modeling helps to represent the various types of data one can use and store within the system.
  • It also helps to represent the various kinds of relationships among these data types.
  • Not just that, but also to represent various ways in which one can group a certain data.
  • To represent various ways in which one can organize a certain data.
  • To represent its various formats and attributes.

Data modeling contributes to the development of internal data standards and uniform data definitions. This is common in the area of data governance projects. It is also important in the data architecture process, which records data assets and maps how data moves across IT systems in order to develop a conceptual data management framework. Data modeling has become an important tool for both data analysts and data scientists involved in the development of BI tools as well as involved in more complicated data science along with advanced analytics.

Data models can divided into various types.

Types Of Data Models

The following are the different kinds of data models:

Conceptual Data Models

A conceptual data model is a pictorial representation that outlines the business ideas that are entities, as well as the interactions between them, in order to learn, showcase, and record a data-driven insight of the organization’s operation. These models are a high-level graphical representation of data of a business. They are not, however, bound to any particular database or application technology. Domain models are another term for this since they provide a big-picture perspective of what the system will comprise. These models are often created as part of the original project specifications’ collection process. One often expresses such models using an ERD (Entity Relationship Diagram) or an ORM (Object Role Model). They either have just important attributes or none at all.

Logical Data Models

It’s time to get more precise using a logical data model after your original concepts have become evident through conceptual data modelling. Logical data models depict the relationships between data items and represent the data from a technical standpoint. These models also clarify the data properties that define such entities, as well as the relationships that are between them.

Basically, a logical data model provides as much information about the data as necessary. This model specifies the primary key for each entity. All foreign keys, or keys that identify the connection between different units, are specified. Normalization also occurs or takes place in a logical data model.

To create such a model, only five steps are necessary.

The steps are as follows:

  • Firstly, you must determine the primary keys for all entities.
  • Secondly, you must determine the connections between various entities.
  • Thirdly, you must identify all of the attributes for each entity.
  • Then, make an effort to repair numerous links/relationships/connections.
  • Lastly, implement normalization.

Physical Data Models

DBAs (Database Administrators) and developers are often the ones that build a physical data model. A physical data model explains how the system would be executed using a certain database management system. This model is  database-specific that depicts relational data items like tables, columns, and relationships between these items. Primary and foreign keys, indexes, authorizations, and so on are all specified in this model. Here, columns should have specific lengths, datatypes, and default values. Database designers take the help of physical data models to build database designs.

The following are some examples of how to make such a model:

  • Using a wizard, you can make a new physical model from a template.
  • You can also import a physical data model file from the file system.
  • You can also convert a logical data model into a physical data model.

There are over four types of physical data models and they are flat-file, hierarchical, relational and object-oriented types.

Understanding The Types

Let us understand these types a little more in detail.

  • Flat-file: A database that is stored in a file is considered as a flat-file. Flat files are divided into two categories, CSV (comma separated values) and delimited files. Both are text-based file formats that store relational data.
  • Hierarchical: According to wikipedia, “A hierarchical database model is a data model in which the data are organized into a tree-like structure.” In this, as the name suggests, data is basically stored in a hierarchical like structure. This model covers parent-child relationships in general, where each parent can have several children but each kid must have a single parent.
  • Relational: Here, relational databases store predicates as a set of constraints, relationships, possible values, and value boundaries. It enables users to submit data and inquiries.
  • Object-Oriented: This model revolves around object-oriented programming. We can build this model using real-world entities, including their properties and links. It detects all objects in the real world that meet the requirements.

Advantages of Data Modeling

The following are few advantages of data modeling:

  • With the help of data models, one can build applications more effectively and efficiently. According to dataversity, It consumes less than 10% of a project’s budget and decreases the 70% of expense allocated to programming.
  • DBAs (Database Administrators) can use data modeling to better understand their databases and optimize them for optimal performance.
  • It reduces software and database development failures.