Paper 2

The Web Within: leveraging Web standards and graph analysis to enable application-level integration of institutional data

Authors: Luiz Gomes-Jr and André Santanchè

Volume 19 (2015)

Abstract

The expansion of the Web and of our capacity of producing and storing information have had a profound impact on the way we organize, manipulate and share data.We have seen an increased specialization of database back-ends and data models to respond to modern application needs: text indexing engines organize unstructured data, standards and models were created to support the Semantic Web, Big Data requirements stimulated an explosion of data representation and manipulation models. This complex and heterogeneous environment demands unied strategies that enable data integration and, especially, cross-application, expressive querying. Here we present a new approach for the integration of structured and unstructured data within organizations. Our solution is based on the Complex Data Management System (CDMS), a system being developed to handle data typical of complex networks. The CDMS enables a relationship-centric interaction with data that brings many advantages to the institutional data integration scenario, allowing applications to rely on common models for data querying and manipulation. In our framework, diverse data models are integrated in a unifying RDF graph. A novel query model allows the combination of concepts from information retrieval, databases, and complex networks into a declarative query language that extends SPARQL. This query language enables flexible correlation queries over the unified data, enabling support for a wide range of applications such as CMSs, recommendation systems, social networks, etc. We also introduce Mappers, a data management mechanism that simplies the integration of heterogeneous data and that is integrated in the query language for further flexibility. Experimental results from real data demonstrate the viability of our approach.