Defining and executing assessment tests on linked data for statistical analysis
Benjamin Zapilko, Brigitte Mathiak · 2011
Abstract. Currently there is a strong trend for governmental agencies to publish statistical data as Linked Data (e.g. Eurostat, data.gov.uk). Unfortunately, these published datasets are still very diverse in their structure, making the analysis very complicated and technical. In this paper, we analyze datasets according to defined assessment tests and exploit both domain knowledge and the inherent semantic annotations. Therefore we scan existing datasets for known patterns that signify e.g., typical numerical data blocks or potentially temporal and geographical dimensions. Thus, Linked Data is made evaluable for a possible usage in standard statistical analysis tools. This allows researchers to use statistical data from diverse Linked Data sources for analysis with only a minimum of technical expertise used for integration of the data.