Paper Demonstration

A substantial effort is being devoted by the scientific community to publishing a large amount of information as Linked Data over the Web. However, the exploration of the Linked Data space is still hindered by the limitations of the current user interfaces, which demand users to manually perform several navigation steps in order to follow the links paths she is interested in. Moreover, data linkage happens at data publishing time, and users are somehow artificially constrained to follow only pre-defined links, even if other forms of data associations could be exploited for navigation. In this paper we propose a novel approach to Linked Data exploration, based on a new class of applications that support the customization of Linked Data search for vertical domains of interest, integration between link navigation and value-based association navigation, and the fusion of search results through join operations across heterogeneous data sources.

Demonstration Goal

This online prototype aims at showcasing a novel approach to Linked Data navigation, providing richer data navigation and exploration through configurable user interfaces, while offering mechanisms to create relationships between Linked Data and non-Linked Data sources. We define a new class of applications that support:

  • Configuration and customization of Linked Data search for vertical scenarios;
  • Integration with label-based search;
  • combination of multiple search results through join operations upon multiple (and heterogeneous) data sources.

Our aim is to increase the accessibility to Linked Data for end users and to lay the path to future commercial exploitation of Linked Data as information sources for vertical application development. In order to achieve such a goal, data discovery, exploration, and consumption must be proposed through data representation paradigms that are familiar to non-expert users, such as pivot tables, filters, facets and re-ordering. To improve the user experience, we propose to support the development of vertical application tailored to specific usage scenarios.

Running example

To describe our approach, we base our discussion on a running example that allow users to search for scientific publications and related information. The user is expected to submit a (part of) conference paper title, the year of publication, and possibly one or more authors; as a response, the systems extracts for him the list of matching papers, of the corresponding authors and conferences, the number of citations of each paper, and the impact factor of the conference. Afterwards, the user can decide to navigate further information about the other papers published in the same conference, the journal papers of each author, and the European research projects in which the authors have worked.

The data sources used in the running example are:

 

Watch the demonstration video