|
Abstract:
|
A long missing feature in R graphics systems is support for interactivity. A number of standalone systems exist, GGobi, MANET and Mondrian, which support interactive displays of multivariate data, but they lack the extensibility, and tight integration with modeling that R furnishes. There have been several attempts to provide interactive graphics in R, from locate() and getGraphicsEvent() in base R, to packages RGtk, RGtk2, rggobi, tcltk and iplots. Iplots is fast and includes most common types of plots; basic operations in interactive graphics. However, it lacks some features such as the tour or direct support for color palettes.
We have developed a new R package, cranvas, which is based on several other packages to make fast (Qt), flexible (R) and elegant interactive statistical graphics in R. The cranvas package aims to borrow from the design of ggplot2, which is based on a grammar of graphics. Currently this package includes common statistical graphics, histogram, scatterplot, bar plot, boxplot, par-coords plot, map and tours, and common interactions, brushing, identifying, deletion, zooming, panning and different types of linking. Color palettes are also supported.
|