Towards Interoperable Open Statistical Data

Abstract

An important part of Open Data is of statistical nature and describes economic and social indicators monitoring population size, inflation, trade, and employment. Combining and analysing Open Data from multiple datasets and sources enable the performance of advanced data analytics scenarios that could result in valuable services and data products. However, it is still difficult to discover and combine open statistical data that reside in different data portals. Although Linked Open Statistical Data (LOSD) provide standards and approaches to facilitate combining statistics on the Web, various interoperability challenges still exist. In this paper, we define interoperability conflicts that hamper combining and analysing LOSD from different portals. Towards this end, we start from a thorough literature review on databases and data warehouses interoperability conflicts. Based on this review, we define interoperability conflicts that may appear in LOSD. We defined two types of schema-level conflicts namely, naming conflicts and structural conflicts. Naming conflicts include homonyms and synonyms and result from the different URIs used in the data cubes. Structural conflicts result from different practices of modelling the structure of data cubes.