Visualization of tabular data—for both presentation and exploration purposes—is a well-researched area. Although effective visual presentations of complex tables are supported by various plotting libraries, creating such tables is a tedious process and requires scripting skills. In contrast, interactive table visualizations that are designed for exploration purposes either operate at the level of individual rows, where large parts of the table are accessible only via scrolling, or provide a high-level overview that often lacks context-preserving drill-down capabilities. In this work we present Taggle, a novel visualization technique for exploring and presenting large and complex tables that are composed of individual columns of categorical or numerical data and homogeneous matrices. The key contribution of Taggle is the hierarchical aggregation of data subsets, for which the user can also choose suitable visual representations.The aggregation strategy is complemented by the ability to sort hierarchically such that groups of items can be flexibly defined by combining categorical stratifications and by rich data selection and filtering capabilities. We demonstrate the usefulness of Taggle for interactive analysis and presentation of complex genomics data for the purpose of drug discovery.
Taggle: Scalable Visualization of Tabular Data through Aggregation
arXiv preprint, 2017.
We thank Bikram Kawan for his contributions to the initial prototype implementation as well as Christian Haslinger and Andreas Wernitznig for providing valuable conceptual feedback. This work was supported in part by Boehringer Ingelheim Regional Center Vienna, the State of Upper Austria (FFG #851460), and the US National Institutes of Health (U01 CA198935).