Map4Scrutiny - A linked Open Data solution for politicians interest registers

Abstract

This article describes the process and results of providing data from the members of the Portuguese Parliament interest registers and the Public Procurement database as Linked Open Data. The organizations mentioned in both link the two resources. The two resources are linked by the organizations mentioned in both. This article focuses on two areas: design and implementation. The first depicts the process of designing an application profile, structuring and describing the data, and creating shapes in ShEx for validation. Both existing and new properties, classes, and controlled vocabularies are used to describe the data in triples. During implementation, OpenRefine is used to clean and uniformize data, reconcile values into links, map triples from tabular to RDF, and export in RDF Turtle. The exported data is validated against the defined ShEx shapes, published in a Triplestore, and queried with SPARQL. The queries are used to showcase the difference between the sourced data and the resulting linked dataset. This article describes every step of the implementation and provides a global vision of how to create Linked Open Data from structured and unstructured sources. These descriptions are particularly useful for researchers and professionals aiming to develop a similar project.