In many domains, it becomes more and more common that an analysis spans various interlinked data sources that we collectively term data landscape. Yet for the selection of appropriate data sources from the wide range of available ones, current approaches and systems rarely offer more support than a File-Open dialog. This paper presents a visualization approach that aims to give a stable and meaningful overview of a data landscape to ease finding and selecting data sources that may be useful in a subsequent visual analysis. As such, it serves as a visual starting point from which to bootstrap a visual analysis by finding and selecting the data sources relevant to a question at hand. This approach is exemplified by applying it to a current snapshot of the data landscape from the CKAN-LOD data hub consisting of 216 data sources with 613 links between them.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below