Graphics Reference
In-Depth Information
he invention (or re-invention) of graphical techniques for discrete and categor-
ical data;
he application of visualization methods to an ever-expanding array of substan-
tive problems and data structures; and
Substantially increased attention to the cognitive and perceptual aspects of data
display.
hese developments in visualization methods and techniques arguably depended on
advances in theoretical and technological infrastructure, perhaps more so than in
previous periods. Some of these are:
Large-scale statistical and graphics sotware engineering, both commercial (e.g.
SAS) and non-commercial (e.g. Lisp-Stat, the R project). hese have oten been
significantly leveraged by open-source standards for information presentation
and interaction (e.g. Java, Tcl/Tk);
Extensions of classical linear statistical modelling to ever-wider domains (gener-
alized linear models, mixed models, models for spatial/geographical data and so
forth);
Vastly increased computer processingspeedandcapacity, allowing computation-
ally intensive methods (bootstrap methods, Bayesian MCMC analysis, etc.), ac-
cess to massive data problems (measured in terabytes) and real-time streaming
data. Advances in this area continue to press for new visualization methods.
From the early s to mid- s, many of the advances in statistical graphics con-
cerned static graphs for multidimensional quantitative data, designed to allow the
analysttoseerelationsinprogressivelyhigherdimensions.Olderideasofdimension-
reduction techniques (principal component analysis, multidimensional scaling, dis-
criminant analysis, etc.)ledtogeneralizations ofprojecting ahigh-dimensional data-
set to 'interesting' low-dimensional views, as expressed by various numerical indices
that could be optimized (projection pursuit) or explored interactively (grand tour).
hedevelopmentofgeneralmethodsformultidimensional contingency tables be-
ganintheearly s,withLeoGoodman( ),ShellyHaberman( )andothers
(Bishop et al., ) laying out the fundamentals of log-linear models. By the mid-
s, some initial, specialized techniques for visualizing such data were developed
(four-fold display (Fienberg, ), association plot (Cohen, ), mosaicplot (Har-
tigan and Kleiner, ) and sieve diagram (Riedwyl and Schüpbach, )), based
on the idea of displaying frequencies by area (Friendly, ).Of these, extensions of
the mosaicplot (Friendly, , )have proved most generally useful and are now
widely implemented in a variety of statistical sotware, most completely in the vcd
package (Meyer et al., ) in R and interactive sotware from the Augsburg group
(MANET, Mondrian).
It may be argued that the greatest potential for recent growth in data visualiza-
tion came from the development of interactive and dynamic graphic methods, al-
lowing instantaneous and directmanipulation of graphical objects and related statis-
ticalproperties.Oneearlyinstance wasasystemforinteracting withprobabilityplots
(Fowlkes, ) in real time, choosing a shape parameter of a reference distribution
Search WWH ::




Custom Search