Conceiving a Multiscale Dataspace for Data Analysis
Matheus Silva Mota, André Santanchè · 2016
A consequence of the intensive growth of information shared online is the increase of opportunities to link and integrate distinct sources of knowledge. This linking and integration can be hampered by different levels of heterogeneity in the available sources. Existing approaches focusing on heavyweight integration – e.g., schema mapping or ontology alignment – require costly upfront efforts to handle specific formats/schemas. In this scenario, dataspaces emerge as a modern alternative approach to address the integration of heterogeneous sources. The classic heavyweight upfront one-step integration is replaced by an incremental integration, starting from lightweight connections, tightening and improving them when benefits worth such effort. Based on several previous work on data integration for data analysis, this work discusses the conception of a multiscale-based dataspace architecture, called LinkedScales. It departs from the notion of integration-scales within a dataspace, and defines a systematic and progressive integration process via graph-based transformations over a graph database. LinkedScales aims to provide a homogeneous view of heterogeneous sources, allowing systems to reach and produce different integration levels on demand, going from raw representations (lower scales) towards ontology-like structures (higher scales).