A Grammar of Graphics: Past,
Present, Future
The grammar of
graphics can describe any static graphic in terms of a series of orthogonal
components. In this talk I will
motivate the grammar, describe my research on the grammar, and how it fits into
my broader research agenda
To give you a feel
for the types of plot that the grammar enables, I'll show examples of
multivariate spatio-temporal plots from a recent
paper, focussing on how the notion of independent
layers facilitates the development of complex graphics. Then, using the grammar, I'll look at
how we can develop new graphical methods that work hand-in-hand with numerical
techniques to improve our understanding of data. In particular, I'll illustrate how we
can add statistical summaries to our plots to develop a sort of visual ANOVA,
and discuss some of the deep connections between graphics and models.
The second half of
the talk will have a more practical bent.
I'll cover:
* how to use qplot instead of base or lattice graphics
* the seven
components of a ggplot plot (the data, aesthetic
mappings, geometric
elements, statistical transformations, position adjustments, scales, the
coordinate system, and facetting)
* combining these components to make
complex, layered, graphics