What is Data Modelling?

Index

To understand data modelling we need to comprehend what is a data model.

A data model is anything that organises elements of data and then relates these elements with each other by forming a relationship. And the process of creating a data model is data modelling.

Why Data Modelling is becoming more popular

Data modelling is essential to support the business operations and user application in the current digital environment. The term globalisation become very popular during the 20th century as people started to migrate to different countries (usually from developing to developed countries). Now in the 21st century a new word has become a phenomenon and that digitalisation.

Every business you can think of is going online and those business who do not go online go out of business and the best example is Blockbuster. Why? Because it has become more apparent than ever for people to go online. Easy of convenience to shop, cheaper prices, more options to choose from and better technology. All the of these factors led people to shop online.

In addition to that, a new wave of disruption came in the name of “Social Media”. People are now more accustomed to using social media platforms to view videos and interreact with people internationally.

More people being online means more data is being created. More data creation means more data can be collected for various reasons. E.g. market research or data analysis.

As more and more data is created we would obviously need to store this data somewhere and here comes database modelling.

Data models are used to organise, store and link data. Thus, great importance is given to data models as it is clinical for the business as it stores importance information about its customers and products.

The process starts with data modelling always starts with “gathering information”. The next step is “conceptional data modelling”. It is basically mapping or writing down how you want to form tables and link them. Conceptional data modelling can be as easy as just writing down what names you want to give to the columns and what information you want in each column (e.g. data type, number of rows) in excel spreadsheet.

The 3rd step would be to implement “logical data modelling” by mapping conceptional data modelling to logical data modelling using the concepts of tables, columns and schemas. We can do that through Normalization. Normalization is the process of normalizing the data by separating the main table into individual tables. Normalization is done to reduce data redundancy and optimise the data structure by grouping the columns appropriately.

Once we mapped the structure, we move to 4th stage “Physical data modelling”. In other words we bring writings to action by creating tables and columns in SQL and then forming a relationship. This is done via Data Definition Language (DDL).

Data Modelling needs to be efficient because in the later stages if an analyst wants to query the data then it should return the information quickly. If the time taken to return a query result is slow then it is a sign of bad data modelling.

So, it is critical to choose which tables will have which specific columns because in the later stages you do not want the analyst to do 4 different joins just to find one information (e.g. customer email).

You also would want to make the tables scalable because in the later stages you would want to add a new column or even a table (that links to this table). Therefore, it is always a good practise to have a primary key or unique column for each table.

The most popular (and easiest) method to organise elements of data and then form a relationship is done via creating tables in a database. A Database are bunch of virtual tables that stores data. These virtual tables are connected to each other via a relationship. This is known as relational data modelling. So, in other words we can refer data modelling to a database modelling. In the diagram below, it shows a typical database model where there are bunch of tables linked via lines based on common columns. A bunch of related table are also known as a “schema”.

The relational database is made and managed digitally via a software system called “Relational Database Management System” (RDBMS). To interact with the system a special computer language is used and the language used to interact with RDBMS is SQL which was created back in the 1970s. SQL stands for Structured Query Language.

Related posts:

When my grandmother died, I wonder if her imaginary friend died with her. Actually, he wasn’t a friend at all. He traveled with her from Znamenka, Ukraine, to Toledo, Ohio, in 1993. He slept on her…

Being raised by narcissistic caregivers or growing up in a toxic environment affects us much deeper than we may realize and can affect our mind, emotions, and body more than we may be consciously…

I wanted to just write a quick update post about how things are going with the Tomorrow People Initiative. The idea was suggested to me by one of the regular members, and I finally had some time to…