E orts towards the wider adoption of Linked Open Data, including the implementation of Linked Data principles and the consumption of Linked Data, are evident in the existing literature and available tools. Yet, these e orts rarely cater for stakeholders who are not familiar with RDF or SPARQL. Hence, we propose the ExConQuer Framework, which facilitates the publication and consumption of RDF in a variety of generic formats. In this manner, any stakeholder can export and work with RDF data in the formats they are most accustomed with, thus lowering the entry barrier to the use of semantic technologies, and possibly enabling the exploitation of Linked Open Data to its full potential.
The extraordinary growth in volume of the Linked Open Data (LOD) Cloud1 is
evidence enough that Linked Data practices are being adopted at an increasing
rate, and accessibility to raw data, especially in recent years, is being given
considerable importance [
by approaches supporting them to consume existing Linked Data in formats
other than RDF [
In this paper we propose the ExConQuer (Extract, Convert and Query) Framework3. This encompasses a number of tools and technologies intended for less experienced or novice users of Linked Data. Through RDF softening (as opposed to semantic lifting) we generate semantically-shallow representation formalisms of RDF data views, whilst retaining the semantic richness of RDF through provenance information. Our aim is therefore to lower the entry barrier to Linked Data reuse, and enable stakeholders to exploit the full potential of Linked Data without requiring to know RDF or SPARQL. 2
The ExConQuer Framework aims to assist stakeholders in consuming and publishing Linked Data. We enable them to explore existing datasets, construct SPARQL queries, generate di erent views of the results, and publish a representation of the undertaken tasks to enable further use. These functions are provided through the Query Builder Tool4, the RDF2Any API, the ConQuer Ontology5, and the PAM Tool6.
Figure 1 shows an overview of the architecture within the framework. The user can explore datasets and create a SPARQL query through the Query Builder Tool, then query a datastore (through a SPARQL endpoint) through API calls. Information pertinent to the executed processes is then persisted in a triple store as Linked Data Publications. Linked Data Publications are automatically published in the PAM Tool once a user downloads the results. Consequently, users can access all the Linked Data Publications through the Provenance-Aware Management Tool (PAM Tool), which allows a user to re-run existing queries, or modify them through the Query Builder. 2.1
Our approach is intended to be particularly user friendly and simple, to allow non-experts to easily use the tool to achieve the goal of re-using open data. Through the RDF2Any RESTful API, the Query Builder Tool enables users to navigate through classes, subclasses, instances, and properties without requiring to know the underlying structure of RDF data. The Query Builder Tool allows users to explore open datasets and, through the execution of a few simple steps, generate a SELECT SPARQL query without requiring any prior knowledge of the query language. The UI therefore enables the user to select a dataset, a class 3 More information on the framework, including source code, can be found here: http: //eis.iai.uni-bonn.de/Projects/ExConQuer 4 http://purl.org/net/exconquer/builder 5 http://purl.org/eis/vocab/cqo 6 http://purl.org/net/exconquer (or a subclass), and properties to include in the results. The user can also add lters to restrict the results even further. After each user selection, the SPARQL query that is generated on the y is updated, so the user also has the option to modify it. Finally, the user can preview a sample of the results, then download the complete result set in one of the provided formats, namely RDF, CSV, JSON, RDB, and a con gurable conversion that is intended for more experienced users. The latter allows a user to convert RDF into potentially any output format, such as XML, KML, TSV (tab separated values) through passing the required parameters in a template. 2.2
RDF2Any API This RESTful API provides the functionality to the Query Builder Tool. The aim of this API is to hide the complexity of RDF and the datasets' underlying schema. The available API calls include getting all classes in a given dataset, getting classes that match a given keyword (in their labels or URIs), getting the subclasses, instances, and properties of a given class, and converting (softening) a result set in a number of given formats. 2.3
All the processes executed through the ExConQuer Framework generate what we call a Linked Data Publication. A Linked Data Publication consists of all the generated information, including the SPARQL query used, its description, the dataset(s) queried, the initial and target data formats, and the user generating the Linked Data Publication instance. We represent all this data using the ConQuer Ontology, which reuses concepts from the SPIN vocabulary7 and the PROV-O Ontology8. The use of PROV-O enables us to represent provenance information. The ConQuer Ontology allows us to replicate the resulting Linked Data Publications and edit them to achieve di erent results (Figure 2). This allows us to implement RDF softening without actually compromising on the the richness of RDF representation, as any resultsets in formats other than RDF are linked back to the original data in RDF. 2.4
We implemented the PAM Tool as a provenance-aware publishing and consumption management tool that enables the exploration of Linked Data Publications through a faceted browser. Through the use of the ConQuer ontology, the Linked Data Publications have queryable metadata that enables users to search for speci c instances using various criteria, such as by the datasets used and the classes queried for. This tool thus allows users to share, explore, directly edit (through the Query Builder or otherwise), and re-use Linked Data Publications, whilst keeping data lineage intact.
The ExConQuer Framework provides solutions that encourage and enable the re-use of open data. Our approach is targeted towards inexperienced users, thus we do not assume that users of our framework are familiar with the Linked Data paradigm. A preliminary evaluation indicated that the tools are intuitively easy to use and useful with regards to their intended use. Through the Query Builder Tool, we aim to lower the entry barrier for any stakeholder requiring the use of Linked Open Data. We enable the user to explore existing Linked Data and generate a SPARQL query, then proceed to download and convert the results in a number of formats. Through the PAM Tool, the user is able to explore existing queries executed on various datasets through lters, and re-load them on the Query Builder tool to edit or re-run them. We showcase all these functionalities in the Demo Video9 (replicable in http://purl.org/net/exconquer/builder).
8 http://www.w3.org/TR/prov-o/ 9 https://youtu.be/ZqS1d0iGcss