When most people think of a Headless CMS, they picture marketing websites or blogs. They're probably not thinking of any combination of words including “federated marine data portals using standardized metadata vocabularies and real-time geospatial services.”

But that’s exactly where Deltares, a Dutch applied research institute for water and subsurface, has pushed the boundaries of what a headless CMS can do.

The project above is based on the building blocks of OpenEarth. These building blocks are set up using off the shelf open source products that are set up to provide and support OGC standards for storing and exchange of spatial data. The show case can be seen live via the Marine Data Store (informatiehuis marien).

At Deltares, a collaborative team of data scientists and engineers under the OpenEarth initiative work to make complex geospatial data FAIR (Findable, Accessible, Interoperable, and Reusable) and openly available. They provide expert guidance to governments, NGOs, and scientific institutions worldwide on leveraging local data applications. The team tackles massive, messy, and often obscure datasets, such as seal migration patterns in the Wadden Sea, digital elevation models of coastal environments, or mussel bed dynamics over time, requiring specialized expertise and tools to manage and analyze.

Yet, DatoCMS sits right at the heart of this system, powering (meta)data management, enabling dynamic search, and giving domain experts an approachable interface to publish, describe, and share their datasets with the world.

The magic happens when DatoCMS meets tools like Mapbox, GeoServer, Geonetwork and the Open Geospatial Consortium (OGC) API standards. But more on that in a bit.

A Decade-Long Journey Toward Open Data

This all started over a decade ago with the OpenEarth initiative. Gerrit Hendriksen, data scientist at Deltares, describes it less like a platform and more like a philosophy.

OpenEarth is about sharing knowledge, not just data, but also the documentation and code that come with it.

From early on, the goal was to follow FAIR data principles: make everything Findable, Accessible, Interoperable, and Reusable. In practice, that meant using global standards like those defined by the OGC—an organization that sets the baseline for how spatial data is served and consumed.

Deltares built their early tools with raw data services: no UI, no metadata frontend, just APIs and documentation. But they soon hit a wall.

People are scared of technology like GeoServer, like PostGIS databases, things like that. We had services, but nobody could work with them. Only technicians could use the platform. We needed something visual, something for end users.

What followed was a collaboration between Deltares and our friends over at De Voorhoede, who worked together on building an exceptional solution to a highly technical problem.

From Clunky XML to a Modern Metadata Layer

The original frontend required XML files to define what data layers were available. That worked. Technically. But it was error-prone, unmaintainable, and unusable by clients. And these clients were serious organizations: Dutch ministries, coastal research institutes, infrastructure agencies, you get the idea.

The team embarked on a ground-up rebuild, recognizing the need for a more robust and scalable solution. To achieve this, they adopted Mapbox for creating rich spatial interfaces, partnered with our friends at De Voorhoede, to tackle complex content structures and web development, and replaced their legacy CMS with a flexible, API-first system. DatoCMS was introduced as the key to unlocking this new approach, providing the team with the agility and scalability required to support their evolving needs and ultimately deliver a more effective and efficient platform for managing complex geospatial data.

We needed something that let us organize our data in a way that our users could actually find and use it. Not just a backend database, something that worked more like a structured content system.

DatoCMS became the place where data stewards define and maintain metadata schemas. It’s where they manage the "table of contents" that tells the frontend what data layers exist, how they’re described, what thematic categories they belong to, and what external sources they're linked to.

And this isn’t a one-way sync. There’s a crawler system built by De Voorhoede, that connects to the OGC WFS endpoints and pulls unique values from datasets: species names, measurement types, geographic locations. Those values are then stored in DatoCMS, making them fully searchable.

One of the most important layers in modeling and describing landscape is a digital elevation model—where are the low parts, where are the high parts? If you visualize it on a map, then everybody knows, of course. You see blue zones, red zones, and things like that.

While the tech stack might sound complex, the biggest lesson here isn’t about tools. It’s about process.

This platform didn’t come from a product manager making a PowerPoint. It came from sitting in rooms with developers, sprinting with our partners, talking to real users, asking them how they want to work and what would actually help them.

There’s no drag-and-drop page builder here. What Deltares and De Voorhoede built is something more powerful: an environment where the backend is highly technical, but the CMS interface is extremely straightforward and usable. Scientists and civil adminstrators, many of whom aren’t developers, can describe, tag, and organize data layers without writing XML or touching code.

Deltares handles the infrastructure: the GeoServer, the metadata harvesters, the data refresh pipelines. Their partners handle the interface and frontend logic. The end clients get something that feels seamless and interactive.

Making Data Searchable and Understandable

Another exceptional addition to this is the ability to search through the data. When your platform has hundreds of layers, each linked to different datasets, models, observations, or measurements, finding what you need can be impossible. Especially if you don’t know what that thing is called in the database.

Deltares solved this by combining structured metadata with real-time API lookups. They pull values from OGC services, convert scientific species names into common names (using the WoRMS API, which Gerrit still chuckles about), and enable multilingual search across hundreds of datasets.

So if someone types "mussels" or “zeehonden” (seals in Dutch), they don’t get a blank screen, they get relevant datasets, linked layers, and metadata descriptions in their preferred language.

“We don’t translate all the data”, Gerrit clarifies. “Besides the fact that you simply can’t translate raster files (because only numbers), a DatoCMS plugin only translates species names from scientific names to common names to enable discovery of layers with observations of species. These harvested names are also stored in the metadata of these specific layers.”

Maintaining Standards in a Federated World

A huge part of this project is standardization. You’ve got external data providers, legacy systems, and ministry mandates. Gerrit admits, “Nobody likes to follow standards.” But they’re essential if you want federated systems to work.

To manage this, Deltares uses vocabularies like AQUO and BODC. The CMS enforces minimum standards so data providers can’t just dump in arbitrary values.

You need to know what 42 means. What is 42? It could be 42 birds, 42 milligrams per liter of nitrate, or whatever. That’s why we need vocabularies and standards.

They’ve even built a metadata harvester plugin in DatoCMS. Instead of manually copying values, data managers can just link to an external metadata record, and Dato pulls it in and makes it searchable.

Why This Works

A lot of people talk about decoupled systems. Very few actually pull them off at scale, across institutions, in a way that’s not a nightmare to maintain.

This works because Deltares didn’t just build a platform. They built a workflow - a continuous process of iteration, feedback, and dialogue between clients, developers, and scientists.

It’s easy to set up a data service. It’s hard to maintain it over time. And it’s even harder to make it usable by people who weren’t involved in setting it up.

It's also worth pointing out that as the CMS, in no way are we "doing all the heavy lifting" here. Instead, we serve as a flexible bridge between the complex data and the users' needs. Dato's role is to provide a robust foundation for managing metadata, content structure, localized descriptions, taxonomy mapping, and other essential functions, all while being versionable, API-accessible, and intuitively usable for non-technical users, letting them focus on their core work without having to get bogged down in technical complexities.

The frontend is snappy and modern, thanks to De Voorhoede’s clean Vue/Mapbox implementation. The backend is stable, with Deltares' ops team handling data freshness, service availability, and long-term maintainability.

Today, the platform supports multiple environmental viewers: Marine Information, WaterInfo Extra, and others, used by public sector institutions across the Netherlands.

And behind it all is yours truly, something most people wouldn't expect to think of to handle this kind of workload for interactive geospatial data.

Summarizing the role of the CMS

At the risk of repetition, DatoCMS is not where Deltares stores its geospatial data. The CMS enables linking datasets living in dedicated servers like GeoServer, enabling accessibility through OGC standards like WMS and WFS. What we provide is a single point of entry to administrate data structure and the metadata, enabling discovery and usability of datasets. It is where data stewards manage what each dataset represents, how it should be categorized, and how it connects to the rest of the system.

Deltares uses DatoCMS’s structured content models to define metadata schemas that match their environmental and scientific standards. These models include fields for layer names, descriptions, thematic tags, source references, measurement parameters, and translations. This makes it possible to manage hundreds of data layers in a way that is consistent, searchable, and understandable for non-technical users.

One of the key features is DatoCMS’s flexibility with plugins and APIs. Working with De Voorhoede, Deltares built a crawler that connects to WFS endpoints and automatically extracts unique values from datasets, such as species names or measurement types. Another plugin harvests metadata directly from external sources like GeoNetwork. This reduces manual effort and keeps metadata aligned with upstream systems.

Because DatoCMS is fully API-first, this metadata is immediately available to the frontend. Search and filtering work reliably across layers, powered by live metadata without the need for custom backend services. Role-based permissions allow data managers to safely handle updates themselves, without relying on developers.

DatoCMS gives Deltares a maintainable, scalable way to bridge the gap between complex geospatial data services and real-world usability. It helps ensure that environmental data is not only open but also understandable and accessible.