Loon: An Interactive and Extendable Statistical Data Visualization Toolkit

• Activity Number: 257
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Graphics
• Author(s): Adrian Waddell* and Wayne Oldford
• Companies: University of Waterloo and University of Waterloo
• Keywords: Interactive Statistical Graphics ; Data Visualization ; Graphical User Interfaces

Abstract

Loon is an open source toolkit for highly interactive data visualization that is written in Tcl/Tk and ported to R and Python. Interactions with plots are provided with mouse and keyboard gestures as well as via command line control and with inspectors that provide graphical user interfaces (GUIs) for modifying and overseeing plots. In this talk, we present loon's scatterplot widget and its related inspector widgets. Some of the implemented scatterplot features are zooming, panning, selection and moving of points, dynamic linking of plots, layering of visual information such as maps and regression lines, custom point glyphs (images, text, star glyphs), and event bindings. Event bindings provide hooks to evaluate custom code at specific plot state changes or mouse and keyboard interactions. Hence, event bindings can be used to add to or modify the default behavior of the plot widgets. Loon's capabilities are very useful for statistical analysis tasks such as interactive exploratory data analysis, sensitivity analysis, animation, teaching, and creating new graphical user interfaces.