Exploring LOD through metadata extraction and data-driven visualizations

Abstract

Purpose – The purpose of this paper is to present a new approach toward automatically visualizing Linked Open Data (LOD) through metadata analysis. Design/methodology/approach – By focussing on the data within a LOD dataset, the authors can infer its structure in a much better way than current approaches, generating more intuitive models to progress toward visual representations. Findings – With no technical knowledge required, focussing on metadata properties from a semantically annotated dataset could lead to automatically generated charts that allow to understand the dataset in an exploratory manner. Through interactive visualizations, users can navigate LOD sources using a natural approach, in order to save time and resources when dealing with an unknown resource for the first time. Research limitations/implications – This approach is suitable for available SPARQL endpoints and could be extended for resource description framework dumps loaded locally. Originality/value – Most works dealing with LOD visualization are customized for a specific domain or dataset. This paper proposes a generic approach based on traditional data visualization and exploratory data analysis literature.