This is the third and final article in our series based on the topic Developing Data-Centric Software. In the first blog, we explored why organizations get stuck in their software landscape and how data-centric working helps. In the second, we looked at why traditional software keeps your data locked in. Now, we dive into the practical side: how can you actually build data-centric software? And what role does SEm® play in making that happen?

Building your own: start with a semantic model

Suppose you want to develop an application that helps your organization become more data-driven. Where do you start? The first step is to describe the scope of your application in a semantic model, a structured representation of reality independent of any specific software. 

It may be tempting to design a complex model, but the smarter strategy is to keep it simple. Because semantic models are flexible and scalable, you can expand them step by step as new needs arise. For example, a simple model might begin with just assets and attributes, and only later include supporting documents or additional layers of detail.

Figure 1: Example of a semantic model

By modelling your scope in this way, you create the foundation for your application´s data model. That means your software stores data in a way that reflects reality, making it easier to exchange and reuse across different systems. However, modelling can also be challenging, often leading to debates about terminology or structure. That’s where a standardized approach helps. 

SEm®: a standardized semantic model

SEm® (Systems Engineering model) provides a standardized software development foundation. It offers both:

1. SEm ontology: a library of basic concepts, attributes, and relationships, aligned with international standards such as ISO 15926 and NEN-EN 17632. This ensures consistency, modularity, and adaptability across diverse applications and industries.
2. SEm framework: rules for how to represent information in semantic models, making them easier to interpret, share, and extend. One example of these rules is the distinction between high-level abstractions (conceptuals) and specific instances in time and space (individual).

These two components together allow for the representation of general concepts and specific assets across their entire lifecycle, from design to operation and maintenance.

In short: SEm® helps organizations avoid reinventing the wheel when building semantic models, while ensuring that the resulting applications are interoperable and future-proof.

From spreadsheets to software: a practical example

For many organizations, the journey toward data-centric working begins with something familiar: a spreadsheet.

Imagine a project manager keeping track of budgets, timelines, and responsibilities in Excel. This works when projects are small, but as they grow, problems emerge: errors creep in, data is duplicated, and only a handful of people can make sense of the file.

Figure 2: Project information managed in Excel

By transforming this spreadsheet into a semantic model, each row and column is given an explicit meaning. “Project” becomes a defined class, “Manager” maps to the Person class in SEm, and standardized relationships clarify how everything fits together. The result: consistent, scalable, and shareable data.

Once the semantic model is defined, implementing it in software becomes much easier. Whether you use a relational database, a graph database, or even a low-code platform like Mendix or Relatics, your data is already structured to support long-term interoperability. That means less time spent reformatting data and more time spent optimizing workflows and building user-friendly applications. 

Conclusions

Becoming data-centric is not just about adopting new tools; it’s about reshaping how you define and use your data. Starting with semantic models allows organizations to develop applications that are scalable, consistent, and aligned with reality. Frameworks like SEm® accelerate this process by providing a standardized foundation, ensuring your applications remain interoperable and future-proof. Instead of waiting for a complete overhaul, begin small. Transform a spreadsheet, define a semantic model, and see how much easier collaboration and decision-making become.