Integration of public procurement data using linked data

Abstract

Listen

Translate article

Linked data is frequently casted as a technology for performing integration of distributed datasets on the Web. In this paper, we propose a generic workflow for data integration based on linked data and semantic web technologies. The workflow comes out of an analysis of the application of linked data to integration of public procurement data. It organizes common data integration tasks, including schema alignment, data translation, entity reconciliation, and data fusion, into a sequence of repeatable steps towards achieving a unified view of the combined datasets. We propose this workflow with the main goal of improving the usability of data and its value for data analyses. Specifically, we target the domain of public procurement and its characteristic challenges this application area poses to data integration. As the paper progresses through the individual steps of the integration workflow, it highlights the key concerns and proposes solutions based on linked data to address these concerns. The discussed data integration methods are illustrated and evaluated on a running example describing integration of public procurement data from the Czech Republic. Using a motivating example of an analytical query on public procurement data we demonstrate how data integration enables to reveal a likely case of clientelism between a contracting authority and its supplier, which is validated by previous coverage in the media.