|
|
|
|
|
What's your data saying? The playitbyr package for
the lovely R statistics & data
environment allows you to listen to a data.frame in
R by mapping columns onto sonic parameters, creating an auditory
graph. Many clever people have developed tools and interfaces for
sonification, but there's just not that much out there to sonify
straight from a statistics
package. Here's how to get
started.
playitbyr 0.2-1 works arm-in-arm with the
venerable Csound synthesizer to bring you wild new sounds:
## All four measurements of the iris datasets, mapped on to pitch, time, modulation frequency, index of modulation
sonify(iris, sonaes(time = Petal.Length, pitch = Petal.Width, mod = Sepal.Length, indx = Sepal.Width)) + shape_scatter() +
scale_time_continuous(c(0, 10)) + scale_pitch_continuous(c(7, 12))
## An audio boxplot: distribution of Sepal.Length in iris dataset
## Pitch varies full range, then interquartile range (25% to 75%),
## then median, for setosa, versicolor, virginica
sonify(iris, sonaes(pitch = Sepal.Length, indx = Species)) + sonfacet(Species) + shape_boxplot(length = 1, tempo = 1800)
## An audio dotplot of the same
sonify(iris, sonaes(time = Sepal.Length)) + shape_dotplot(jitter = 0.3) + sonfacet(Species)
To make sonification easier to learn, the package mimics the syntax
of the popular and beautiful
ggplot2
package. sonify objects, created by
the sonify() function, contain both the data to be
sonified and the mappings and scalings of the sound
parameters. Users can tweak their sonification by adding other
mappings, scalings, facets, and data to the mix.
One of the first experiments in sonification for statistical exploration was at Bell Labs in 1977, a collaboration which included John Chambers and Max Matthews. John Chambers created S, and Max Mathews worked on the groundbreaking MUSIC. Isn't it fitting that their open-source children, R and Csound, should decide to cozy up once again?