Skip to main navigation Skip to search Skip to main content

ConnectedCharts: Explicit visualization of relationships between data graphics

  • École de technologie supérieure

Research output: Contribution to journal › Journal Article › peer-review

25 Citations (Scopus)

Abstract

Multidimensional multivariate data can be visualized using many different well-known charts, such as bar charts, stacked bar charts, grouped bar charts, scatterplots, or pivot tables, or also using more advanced highdimensional techniques such as scatterplot matrices (SPLOMs) or parallel coordinate plots (PCPs). These many techniques have different advantages, and users may wish to use several charts or data graphics to understand a dataset from different perspectives. We present ConnectedCharts, a technique for displaying relationships between multiple charts. ConnectedCharts allow for hybrid combinations of bar charts, scatterplots, and parallel coordinates, with curves drawn to show the conceptual links between charts. The charts can be thought of as coordinated views, where linking is achieved not only through interactive brushing, but also with explicitly drawn curves that connect corresponding data tuples or axes. We present a formal description of a design space of many simple charts, and also identify different kinds of connections that can be displayed between related charts. Our prototype implementation demonstrates how the connections between multiple charts can make relationships clearer and can serve to document the history of a user's analytical process, leading to potential applications in visual analytics and dashboard design.

Original languageEnglish
Pages (from-to)1285-1294
Number of pages10
JournalComputer Graphics Forum
Volume31
Issue number3 PART 4
DOIs
Publication statusPublished - 2012

Fingerprint

Dive into the research topics of 'ConnectedCharts: Explicit visualization of relationships between data graphics'. These topics are generated from the title and abstract of the publication. Together, they form a unique fingerprint.

Cite this