Data Modelling Skills: A Must-Have for Data Success

Data modelling is often pushed to the side, thought of as a specialist or technical skill, but it’s a fundamental part of managing data well. Everyone from the data modeller to the data analyst should have data modelling skills – in this article, we explain how this delivers value.  

What is data modelling?  

In simplest terms, data modelling is about planning your data. The best analogy is to compare it to building a house. Would you build a house without seeing the floor plan and elevations? Of course not. Therefore you shouldn’t be designing databases or adding new technology without data models in place.  

While definitions are often dismissed as ‘just semantics’, they are key to creating a common understanding. Data modelling maps out the core definitions and business rules on which applications, databases and analytics that run an organisation are built. It establishes stable data foundations that provide durability – so that technology is robust enough to withstand major changes. 

Data modelling defines core terms like: 

• What is a customer?
• What is a product, an order, or a region?  

While these terms may seem straightforward, it’s common for different departments and individuals to define them differently. Data modelling also maps the relationships between these things. For example A customer may place many orders over time. An order may be for one or more products. 

Data modelling uses different levels of abstraction. A conceptual data model shows the big picture, a logical model describes the data elements visible to users and a physical data model is the most detailed, with technical requirements that inform the building of a specific database. 

When is data modelling used?  

Most people turn to data modelling when there’s a greenfield project coming up. A shiny new database or data warehouse to design and develop. 

But data modelling should be part of everyday business. Data modelling should come into play when you: 

• Add a new data source to an existing data warehouse. You’d use a physical data model to understand what’s in the new database and map it to the warehouse. 
• Plan for a new report or dashboard. It’s best practice to use data modelling to document requirements to pass to a business analyst to create the reports and dashboards. 
• Migrate data to the cloud. It’s important not to leave data modelling to the software vendor: only you know your business and your data inside out. 

• Plan for a merger or acquisition. The tricky task of integrating the data from two organisations relies heavily on data modelling. 
• Plan for a change of business strategy. If you already have data models in place, you’re in a much better position to capture new requirements and accommodate change. If you’re starting from scratch, it’s a much longer and costlier job. 

Why is data modelling important? 

If we go back to the house example, you can only imagine the issues with building a house without the owner reviewing the plans. It’s highly unlikely that you’d be happy with the result, and your poor builder would be smashed with variations and rework! 

Same goes for building databases and data warehouses.  

Data modelling creates a common mental model that brings your business and technical teams on the same page. It helps to:  

• Keep costs down. Without solid planning and stable data structures, you are likely to encounter more re-work, systems that don’t work well together, gaps and duplications in your data, confusing reports – and more. 

• Ensure better, faster database design. With data models and business requirements carefully constructed at the outset, your developers can work more efficiently and confidently. Data modelling doesn’t just define requirements, it encapsulates the business rules around data. A bonus is that data modelling helps to break down the barriers between your IT and business teams.  
• Provide a solid base for reporting and analytics. Your dashboards are only as good as the data that feeds them. Data modelling helps to reduce data duplication, enhance data quality and ensure the simplicity of your data structures.  

Who needs data modelling skills? 

Data modelling skills are essential in every organisation. And it’s not just data modellers who need them. Here’s a rundown of who needs data modelling skills and why – it doesn’t mean they all need to be able to create data models – some are more about being able to understand and interpret them. 

• Data modellers: To state the obvious, data modelling skills are essential for data modellers and database designers. What’s important here is to ensure that anyone in a data modelling position understands international best practice (the DMBoK). It’s also the case that many data modellers won’t have studied modelling in their tertiary studies, especially if they didn’t set out to be a data modeller. 
• Data architects: Data modelling is a must-have in any data architect’s toolbox as this discipline shares responsibility for data planning – just at a higher level. Understanding data modelling gives data architects a concrete idea of what data modellers require from them. This in turn ensures that architects provide enough direction and detail for the data modeller to effectively map out databases and data warehouses.  
• Data integration specialist: Understanding data modelling makes the data integrator’s life easier. They can model data in two source systems and then design a new data model for the combined data. In short, data modelling aids the data integration process and makes planning easier. 
• Metadata specialists: Data models are a form of metadata – it’s useful for metadata specialists to at least be able to understand and read a data model.  
• Master data managers: Must know how to read and interpret data models. They also need to create the data model for the shared master data repository and read models of all the source systems. 

• Data analysts and data scientists: Data modelling is under-represented in the world of data analytics and data science. It’s a useful skill to be able to interpret data models when creating dashboards and ensuring the reliability of information. When drawing from a data warehouse, it’s a great help to understand exactly how the data is structured. 
• Data warehouse designers: Use dimensional modelling and data vault modelling to design data warehouses. A solid general understanding of data modelling is also required so they can read the data models of the source systems. 
• Subject matter experts: It helps if your SMEs can read high-level logical data models. This aids the communication between the business and IT and helps ensures your projects run smoothly. 

What makes a good data modeller?  

A good data modeller is highly analytical and detailed. Their aim is to create accurate data models. But they also need to be good communicators, so they can understand business requirements around data and ensure the data models reflect this. 

Where can I learn data modelling skills?  

Robinson Ryan’s Institute of Data Management runs a Practical Data Modelling course that is suitable for anyone who needs to create and/or understand data models.  

Training is aligned with DAMA’s Data Management Body of Knowledge – DMBoK V2 – which ensures you are drawing on international best practice. You’ll also be learning from one of only four Registered Education Providers with DAMA International and the only one based in Australia or Asia-Pacific.