In today’s digital world, seamless data exchange and smooth software integration are more important than ever. Two key technologies that aim to tackle these challenges are the Semantic Web and Linked Data (SWLD).

This is especially relevant in industries like Architecture, Engineering, and Construction (AEC), where companies frequently collaborate on complex projects with different teams and partners. In such scenarios, having a common way to share and interpret data becomes crucial.

In this blog series, we’ll explore the key challenges preventing wider adoption of ontologies and linked data within the AEC industry. We’ll also look into how ETL (Extract, Transform, Load) tools can help address these challenges.

Linked Data and Ontologies

Ontologies are gaining attention as a key technology for improving data sharing and interpretation across industries. They provide a common framework that enhances productivity and streamlines information management. Despite the extensive research in this field, particularly in creating and applying ontologies, their real-world implementation still faces challenges.

One major hurdle is integrating ontologies into existing software systems, which often lack the flexibility to handle the complex data structures that ontologies present. This is particularly relevant in the Architecture, Engineering, and Construction (AEC) industry. This is because tools like Building Information Modeling (BIM) struggle to fully support ontology integration.

Semantic Web and Linked Data (SWLD) technology plays a crucial role in publishing and sharing ontologies because it allows data to be machine-readable and clearly defined. The idea is that software can easily access these ontologies and leverage their unified vocabularies and data structures. However, some challenges remain.

One challenge is the variety of modeling patterns that can be used to represent the same data, even within standard vocabularies like RDFS, OWL, and SHACL (read more about these concepts here).
Another challenge is the limited out-of-the-box support for linked data in most applications, which often requires custom development. This makes it harder for organizations to fully take advantage of the potential of linked data and ontologies in areas like 3D modeling and enterprise applications.

ETL

ETL (Extract, Transform, Load) tools are commonly used to integrate data from various sources into a central system. While many studies explore how ontologies and linked data can improve ETL processes, we see it the other way around: ETL tools can help organizations better use ontologies.

These tools can extract data from RDF sources (the foundation of ontologies) and transform it to meet the needs of different applications. This makes it easier for traditional systems to work with linked data, enhancing functionality and promoting better data interoperability.

By bridging the gap between complex ontological frameworks and everyday software, ETL tools help the AEC industry unlock the full potential of linked data. Ultimately leading to more efficient operations and smarter information use.

Plug-in Framework

To reduce the dependence on software vendors for specific interoperability solutions, we propose a generic plug-in approach that leverages semantic web and linked data technology. This method works with any existing linked data ontologies. As well as avoids the need for vendor-specific implementations. Given the limited native support for linked data in most applications, some form of middleware or plug-in is necessary to enable this functionality. Our proposed framework outlines four key capabilities that can be used to integrate parts of linked data ontologies into existing software providing, a flexible solution to unlock the benefits of linked data without waiting for native application support.

Image 1: Plug-in framework

These capabilities can be described as:

1. Referencing: This is the simplest method for reusing ontology knowledge. Instead of importing the actual ontology content, the application shows a view of an external ontology and stores references to its identifiers. This allows data in the software to be linked to ontologies, such as classifying assets using a reference data library.
2. Importing: This involves loading parts of an ontology, like asset types and definitions, directly into the software’s database. Once imported, this knowledge can be used within the software’s regular interface, enhancing data management and functionality.
3. Structuring: The most ambitious approach involves importing an ontology’s structure into the application’s database. Low-code platforms can sometimes handle this through configuration imports, while traditional systems may require creating new tables or altering the database structure. This method allows the software to function as a direct implementation of the ontology.
4. Exporting: To fully leverage ontology-enriched data, enabling data export in a format that aligns with linked data standards is important. This allows organizations to share and combine their application data with the ontology, ensuring a consistent, interoperable data exchange.

The four capabilities we’ve discussed—referencing, importing, structuring, and exporting—each requires different types of plug-ins and functionalities to enable software to work effectively with linked data ontologies.

For referencing and importing, the key challenge is translating linked data into a format the application can understand. Another challenge is providing a user interface to help domain experts select the relevant ontology content. Our research will focus on these two capabilities.

Structuring an application’s underlying database is the most complex. It also has the greatest impact on existing software, which makes it more difficult to adopt. Due to these challenges, this capability was not developed in our current work.

Finally, while exporting data back into a linked data format is crucial for achieving full interoperability, this aspect falls outside the scope of this research. The primary focus remains on making linked data more accessible and practical through referencing and importing capabilities.

Conclusion

Stay tuned for the other parts of this blog series. There, we’ll dive deeper into how these plug-in capabilities can transform data management in the AEC industry. If you have any questions, please contact us here or reach out to our consultant, Utku Sivacilar.