Revisiting Bertin Matrices: New Interactions for Crafting Tabular Visualizations

Abstract

This is the accepted version of the paper.This version of the publication may differ from the final published version. Abstract-We present BERTIFIER, a web app for rapidly creating tabular visualizations from spreadsheets. BERTIFIER draws from Jacques Bertin's matrix analysis method, whose goal was to "simplify without destroying" by encoding cell values visually and grouping similar rows and columns. Although there were several attempts to bring this method to computers, no implementation exists today that is both exhaustive and accessible to a large audience. BERTIFIER remains faithful to Bertin's method while leveraging the power of today's interactive computers. Tables are formatted and manipulated through crossets, a new interaction technique for rapidly applying operations on rows and columns. We also introduce visual reordering, a semi-interactive reordering approach that lets users apply and tune automatic reordering algorithms in a WYSIWYG manner. Sessions with eight users from different backgrounds suggest that BERTIFIER has the potential to bring Bertin's method to a wider audience of both technical and non-technical users, and empower them with data analysis and communication tools that were so far only accessible to a handful of specialists. Permanent repository link

Fair Statistical Communication in HCI

“…Some empirical data -especially categorical or ordinal data like questionnaire responses -is also hard to faithfully capture with a single aggregated measure. As an alternative, such data can be conveyed without loss of information using compact visualization methods such as matrix displays (Perin et al, 2014). Showing individual observations next to error bars can also be informative and pedagogical (Drummond and Vowler, 2011;Ecklund, 2012).…”

Exploring the Possibilities of Embedding Heterogeneous Data Attributes in Familiar Visualizations

“…Bertin's physical implementation was later adapted to computers [14,47,51], e.g., TableLens [41] and FOCUS [48]. Bertifier [38] is a recent interactive tabular visualization of heterogeneous data that remains faithful to Bertin's interactive matrix technique, while taking advantage of the power of interactive computers. Tabular visualizations of heterogeneous data have usually been considered as independent visualization techniques.…”

“…The HEDA matrix as inspired by Bertin's matrices [5] and the interactive Bertifier [38], is a visualization component that can be integrated into other visualizations. TimeSpan [30] used this technique with stacked bar graphs to make a hybrid visualization that shows stroke patients' multi-dimensional data within the context of their temporal data.…”

“…Reordering rows and columns of a matrix is a known way to reveal patterns in data [38]. As HEDA is a visualization primarily based on this matrix representation, we provide 1) an interactive reordering of attribute locations supporting choice-based comparisons, and 2) a system response to complete the reordering of entities based on one or more currently selected attributes.…”

“…Through application of HEDA we attempt to synthesize the benefits of these alternative approaches. HEDA was inspired by the Bertin matrix [5], its interactive counterpart Bertifier [38], and TimeSpan by Loorak et al, which described how stacked bar graphs could be extended with tabular visualizations (e. g., HEDA) to create a hybrid layout [30]. TimeSpan demonstrated the usefulness of HEDA for exploring the heterogeneous multi-dimensional data of stroke patients.…”