One of the advantages of applying Linked Data practices is ease of decentralized data integration, which, however, is typically an advantage that happens at the backend of applications, and is rarely visible for an end-user. Moreover, we see a rise in applying Linked Data practices for permissioned data sharing, specifically in an industrial setting. To demonstrate the value of Linked Data practices in an industrial setting, we present “Miravi”: a Web application to manage SPARQL queries and visualize their results, integrating data over decentralized heterogeneous Linked Data sources, both openly accessible and permissioned. Stemming from (and fulfilling) the requirements identified within Horizon Europe project Onto-DESIDE, Miravi demonstrates how Linked Data practices, even for permissioned data publishing, allow for ease of decentralized data integration, combining mature and well-supported open-source tools such as Inrupt's Solid Client Authentication library, IDLab's Comunica, and Marmelab's React-Admin Web framework. Miravi is currently used as part of demonstrations for projects such as Onto-DESIDE and SecuWeb, and as a dedicated search interface for Digital Flanders' OSLO program (Open Standards for Linked Organizations). Being a usecase-independent client-side application that is fully configurable with a single configuration file, and under the open MIT license, Miravi allows for a lightweight setup, whilst relying on mature software libraries helps to increase its longevity. With Miravi, we will continue to more easily demonstrate the advantages of Linked Data practices.
One of the advantages of applying Linked Data practices is ease of integrating data across decentralized (and heterogeneous) data sources. This leads to data federation: where traditional data applications require extensive data processing to align diferent sources’ data models and data values, the use of Semantic Web standards allows to integrate data on-the-fly using a single (SPARQL) query. However, this advantage is typically noticeable at the backend of a (Linked Data) application, and is rarely visible for an end-user. To prove its value, Linked Data consumption platforms over decentralized data sources (i.e., Linked Data viewers) are needed.
Moreover, we see a rise in applying Linked Data practices for permissioned data sharing (i.e., publishing Linked Data behind authentication/authorization). This is specifically true in an industrial setting: the need to privately share data with and integrate data from trusted third parties. This adds requirements to a Linked Data viewer, as data is not only published decentralized, but also heterogeneously: both openly accessible and permissioned data can be integrated, using diferent authentication schemes, over diferent interfaces (e.g., publishing a single resource or via a SPARQL endpoint).
To demonstrate Linked Data’s value of easy integration across decentralized, heterogeneous, and permissioned data sources in an industrial setting, we present Miravi (v2.0.0): a Web application to (i) as a Semantic Web developer, manage queries; and (ii) as a non-technical user, visualize query results. Stemming from (and fulfilling) the requirements identified within Horizon Europe project allow for ease of decentralized data integration (specifically for non-technical users), combining mature
CEUR
ceur-ws.org and well-supported open-source tools. Miravi is published at https://github.com/SolidLabResearch/ miravi-a-linked-data-viewer/releases/tag/v2.0.0, under MIT License.
After discussing the identified requirements and matching them with related works (Section 2), we describe the architecture and implementation (Section 3), and the use cases Miravi has been applied to (Section 4). Finally, we conclude (Section 5).
In the Ontology-based Decentralized Sharing of Industry Data in the European Circular Economy
project (Onto-DESIDE)1, semantic ontologies and Linked Data are combined with the concept of
decentralized data sharing to enable data collaboration in the context of circular economy [
The Open Circularity Platform serves as the technical backbone for Onto-DESIDE [
Extended use of Linked Data practices from open data sharing to permissioned data sharing (i.e.,
requiring authentication and authorization) are exemplified by standardization eforts from, e.g., the
Solid project2 and the Fedora Repository3, which both extend the W3C recommended Linked Data
Platform (LDP) [
Within Onto-DESIDE, permissioned data sharing was implemented via the Solid project, which
introduces an extension to the OpenID-Connect (OIDC) protocol [
To demonstrate the value of the Open Circularity Platform (and thus, of Linked Data for easy data integration) to non-technical users, a user-friendly application was needed. To meet this need, Miravi was developed. The following requirements were identified for the development of Miravi, derived from user stories formulated by Onto-DESIDE industry partners6:
Authenticate [R1] Initiate a secure session to interact with your and other actor’s data, behind a layer of access control.
Query [R2] Retrieve specific data from one or more actors in the network. This data is thus by nature decentralized, and can be a combination of permissioned and public data.
Validate [R5] Validate that the retrieved data is genuine and has not been tampered with. Scalability [R6] Scale the management with the number and the heterogeneity of sources and queries. Fault tolerance [R7] Have redundancy by design, so that a failing node does not impact the entire ecosystem. 1https://ontodeside.eu 2https://solidproject.org/, taken up by the W3C Working Group on Linked Web Storage: https://www.w3.org/2024/09/ linked-web-storage-wg-charter.html 3https://fedorarepository.org/ 4https://graphdb.ontotext.com/documentation/10.7/access-control.html 5https://help.metaphacts.com/resource/Help:Security 6https://ontodeside.eu/results/, Deliverable 2.2 “Project requirements specification and research methodology”.
Reproduce [R8] Provide well-documented (open-source) code and APIs.
Multiple Linked Data viewers exists. Examples are collected at, e.g., https://linkeddata-89b9d.web.
app/, and compared in several surveys [
In this section we describe the architecture and implementation of Miravi. We link the supported features to the requirements listed in Section 2.
As an architecture covering the requirements, we present a client-side Web application without serverside processing. This simplifies the overall technological setup: as all processing is done client-side, no additional component can fail. This does exclude the possibility to move any heavy processing to a server-side component and could thus impact overall performance of the application.
To implement this Web application, we maximally rely on mature and well-supported open-source tools such as Inrupt’s Solid Client Authentication library, IDLab’s Comunica, and Marmelab’s ReactAdmin Web framework. We develop the application in a single programming environment, using the Javascript programming language and packaging for browser usage with Vite.
To access permissioned data on Solid Pods, you must authenticate [R1] as a user who has been granted appropriate access to that data. We applied Solid-OIDC, being the Solid authentication protocol. Miravi reuses Inrupt’s Solid Client Authentication library7 to authenticate the users in a browser, producing a custom fetch function, which you can use to execute authenticated HTTP requests.
To retrieve and manipulate the data, we applied the SPARQL Query Language (SPARQL) [
Existing Linked Data viewers primarily rely on a single (SPARQL) endpoint. To support a decentralized ecosystem, we want to support multiple endpoints, scalable to heterogeneous interfaces [R6]: we can not assume that each actor will set up a (resource-intensive) SPARQL endpoint to share data, instead, we assume that heterogeneous data sources are made available, of which we primarily focused on RESTful Web APIs, e.g., the Linked Data Platform API as integrated in the Solid protocols.
Several query engines that support federated SPARQL querying exist [
In Miravi we forward the custom fetch function – produced by the Inrupt library – to Comunica to make it perform authenticated queries over Solid pods and other Linked Data sources. 7https://docs.inrupt.com/developer-tools/javascript/client-libraries/authentication/ 8https://comunica.dev/
CUSTOM QUERIES
SOURCE INSPECTION
EXPORT AS CSV PREDEFINED QUERIES
QUERY RESULT
Comunica has an online user interface9, targeting users with knowledge of Linked Data and SPARQL. Miravi adds more features (Section 3.2), specifically supporting non-technical users. Its user interface (Figure 1) is built with Marmelab’s React-Admin10, an open-source framework for building Web data dashboards, in which we wrap the authentication and query capabilities of Miravi. The configuration of the interface is fully customizable using a single configuration JSON file. Titles, descriptions and icons can be adapted, but mostly the queries can be predefined, grouped in categories.
To validate that the retrieved data is genuine and has not been tampered with [R5], we applied W3C recommended Verifiable Credentials 11.
At its base, Miravi is a data dashboard that integrates data from decentralized (dynamically discoverable) data sources, triggered by underlying configured SPARQL queries (or query templates), to visualize the results in a tabular format.
Miravi allows – for each configured query – to set a predefined set of data sources to query. However,
to avoid the need to configure each data source individually, Miravi supports data source discovery: we
allow to start from a single index source, and then apply Comunica’s link traversal functionality [
Miravi allows specifying preconfigured SPARQL queries. However, to avoid the need to configure each query for each instance separately, we can apply query templates that contain parameters. Before a SPARQL query is submitted, each template variable is dynamically replaced with the actual value selected by the user. The list of available values for each variable can either be predefined or are the result set of another query, i.e., dynamically constructed from data. This approach allows queries to be customized per data type, even when new data of the same type is added.
By combining data source discovery and query templates, the configuration becomes scalable [R6]: queries do not need to be adapted when more data sources and more data are added to the network. 9https://query.comunica.dev/ 10https://marmelab.com/react-admin/ 11https://www.w3.org/TR/vc-data-model-2.0/
The predefined queries enable views and derived calculations [R3-4] on the data sources (Figure 1, left), understandable for users without any knowledge of SPARQL or Linked Data (Figure 1, bottom center). Alternatively, users can also create and share their own custom queries in the interface, resulting in additional custom views (Figure 1, top left). Additionally, users can export query results as CSV files for further analyses and more complex calculations (Figure 1, top right). The user can also inspect the protection, fetch and verification status of the consulted data sources (Figure 1, top center). The verification status validates [R5] any Verifiable Credentials wrapping the data to assure the retrieved data is genuine and has not been tampered with.
Mirava is published on GitHub at https://github.com/SolidLabResearch/miravi-a-linked-data-viewer under the open MIT-license, and with extensive documentation, including a screencast demonstrating its main features, and possibilities to register issues, making it reproducible and reusable [R8].
At the moment of writing, customized instances of Miravi have been set up for three projects: the OSLO Knowledge Graph, Onto-DESIDE, and Secuweb.
Open Standards for Linked Organizations (OSLO)12 is a project of Digital Flanders, that has published at the moment of writing 251 vocabularies and application profiles, at https://data.vlaanderen. be. Miravi was used to publish a dedicated search interface over the published vocabularies and application profiles, available at https://w3id.org/imec/knows/oslo-kg-viewer, as an aid for developers of future standards as well as users of existing standards. Within this setup, data is queried from a small13 set of public raw Turtle data files published directly from Github, with a small set of query templates.
Onto-DESIDE is a Horizon Europe project on data sharing for increased circular economy. Access control is an important requirement for data sharing in this context. The development of Miravi was a part of the Onto-DESIDE project, and enabled the setup of an Onto-DESIDE viewer, available at https://w3id.org/imec/knows/onto-deside/viewer. Within this setup, the integration of a medium14 set of permissioned decentralized Linked Data sources is demonstrated, shared under Solid-compliant access control.
A similar demonstrator was set up for Secuweb15, an Interreg Europe project, aiming to strengthen data security in companies and organizations in Flanders, Wallonia, and France. With Miravi, a Secuweb viewer was instantiated, integrating decentralized Linked Data from the Food Supply Chain domain, with access control, screencast available at https://www.youtube.com/watch?v=CgCVOqOTFFg.
Being a usecase-independent client-side application that is fully configurable with a single configuration ifle, Miravi allows for setting up lightweight Linked Data dashboards that support decentralized and heterogeneous data sources, both open and permissioned (currently supporting Solid-compliant access control). Data source indexes and SPARQL query templates are applied to easily scale the number of data sources and query types. The user interface facilitates access to preset queries for non-technical uses and ad hoc configuration of custom queries for administrators. Query results are presented in a tabular format and can be exported to CSV to prevent lock-in.
Miravi is usecase-independent and flexible, as evidenced by the three current setups for the OSLO Knowledge Graph, Onto-DESIDE, and Secuweb.
Miravi is built on mature software libraries with extensive documentation and published on GitHub under the open MIT license, helping to increase its longevity. Future work includes improving Miravi’s 12https://www.vlaanderen.be/digitaal-vlaanderen/onze-diensten-en-platformen/oslo 13Queries integrate on average 2 data sources. 14Queries integrate on average 70 data sources. 15https://www.interreg-fwvl.eu/nl/projecten/secuweb current JSON configuration to a semantic (JSON-LD) configuration, to improve internationalization support and make query template reusable outside of the configured project.
With Miravi, we will continue to more easily demonstrate the advantages of Linked Data practices.
The described research activities were supported by SolidLab Vlaanderen (Flemish Government, EWI and RRF project VV023/10), and the European Union’s Horizon Europe research and innovation program under grant agreement no. 101058682 (Onto-DESIDE).
The author(s) have not employed any Generative AI tools.