Building Smarter Maps: How Semantic Web Tech Is Transforming Environmental Data Integration

Redacción HC
13/03/2025

Environmental data—think weather sensors, land use maps, river levels, and urban zoning—plays a vital role in everything from flood prediction to sustainable urban planning. Yet across the world, a persistent problem hampers its effectiveness: data fragmentation. These datasets often exist in incompatible formats and isolated systems, making it difficult for decision-makers to access the full picture.

A recent study by researchers at the Norwegian University of Science and Technology (NTNU) and Kartverket (Norway’s national mapping agency) proposes a bold solution rooted in semantic web technologies. Published in ISPRS International Journal of Geo-Information in January 2025, the study outlines a framework that allows for seamless integration and querying of heterogeneous environmental geospatial data using ontology-based data access (OBDA).

This blog post unpacks how the proposed framework works, why it matters, and how it could support smarter and more transparent decision-making in environmental management globally.

Why Environmental Data Doesn’t Always “Talk”

From shapefiles and GeoJSON to real-time sensor APIs and legacy SQL databases, environmental data comes in a dizzying array of formats and vocabularies. Municipalities, research centers, and national agencies often build data systems independently, leading to semantic mismatches and technical silos.

This lack of interoperability isn’t just a technical nuisance—it’s a policy bottleneck. For example, imagine trying to coordinate an emergency flood response using separate datasets from weather stations, urban infrastructure maps, and river gauge sensors that can’t be easily combined or queried together.

The NTNU research team’s core question was:

How can semantic web tools and ontology-based querying improve integration and interoperability of such fragmented data ecosystems?

A Modular Framework That Connects the Dots

The researchers designed a modular framework composed of three main layers:

1. Data Storage Layer

• Uses PostgreSQL with PostGIS, a well-known open-source spatial database system.
• Stores both static data (e.g., building footprints, administrative boundaries) and real-time data (e.g., water level measurements).

2. Ontology-Based Access Layer

• Powered by Ontop, an open-source OBDA platform.
• Translates SPARQL queries (semantic queries) into SQL commands, enabling users to interact with a virtual knowledge graph that maps real-world geospatial concepts.

3. Query Interface

• Built with Leaflet and SQI (SPARQL Query Interface).
• Provides users a web-based interface where they can write SPARQL queries and view results in both map and table formats.

Real-World Test: Lake Mjøsa, Norway

To test the framework’s feasibility, the team integrated six real-world datasets from the Lake Mjøsa region, combining static maps with live environmental sensor feeds. The prototype allowed users to, for example, identify which municipalities were experiencing high water levels—without needing to know the underlying database structure.

Key Findings: Practical Wins for Interoperability

Unified Data Access Without Manual Transformation

Thanks to OBDA and the extended GeoSPARQL ontology, users could query across diverse datasets without manually converting or restructuring them. For example, a query like *“Which weather stations are located within X municipality?”* became possible through one semantic query.

Integration of Real-Time and Historical Data

The framework accommodates dynamic API feeds alongside archived spatial data, offering a low-cost, semi-real-time monitoring system—what the authors refer to as a “lightweight digital twin” for the lake region.

Fully Open-Source and Scalable

By using only open-source tools like PostgreSQL, Ontop, and Leaflet, the framework demonstrates that high-quality semantic integration doesn’t require proprietary systems. This approach is transferable and cost-effective, especially for public institutions.

“This framework acts as a universal translator for environmental data,” the authors explain, highlighting its ability to connect different formats and terminologies into a common querying space.

Why This Matters: Implications for Policy and Practice

This framework isn’t just a research exercise—it’s a blueprint for real-world implementation.

For Governments and Municipalities

Agencies can use this system to centralize fragmented data across departments, allowing for quick decision-making during environmental emergencies or urban planning efforts. For instance:

• Query flood-prone zones by combining river gauge data and building maps.
• Monitor pollution levels using weather station outputs overlaid with population density maps.

For Transparency and Civic Engagement

The open-source, web-accessible nature of the framework empowers citizens, journalists, and NGOs to access meaningful data insights—without needing technical expertise in GIS or SQL.

Imagine a portal that lets any resident visualize rising river levels in real time—linked with demographic risk zones. This system makes that possible.

Recommendations from the Authors

1. Add advanced spatial functions like buffer zones and overlays.
2. Test in higher-volume environments and different geographic regions.
3. Adapt for other domains, such as urban mobility, disaster response, or watershed management.
4. Move toward full-scale digital twins, where the virtual model of an ecosystem dynamically reflects reality.

Conclusion: Turning Data Silos into Smart Systems

This research marks a significant advance in environmental geoinformatics. By blending OBDA, semantic technologies, and open-source tools, the NTNU team has provided a scalable and adaptable solution for one of the most pressing challenges in environmental data science: interoperability.

If you're a policymaker, data engineer, or sustainability advocate, this is your call to action. Start looking beyond formats—and start connecting your data through meaning. The tools are here. The framework exists. The next step is implementation.

Topics of interest

Technology