Spotting Differences - Data Points: Visualization That Means Something - page 252

weak rebounder, and fouled a lot. A line in the parallel coordinates plot for
that player would start high, go low, and then come back up again.

When you plot more units, this method helps to spot groups and tenden-
cies. In the following example, you can use parallel coordinates on data
from the National Center for education Statistics.

CrEATE A PArALLEL CoorDINATES PLoT
You have several interactive options for parallel coordinates. You can
build it in Protovis if you want to make a custom graphic, or you can plug
your data into an exploratory tool such as GGobi. These implementations
enable you to filter and highlight the data points you're interested in; how-
ever, I still like to go with static parallel coordinates plots, namely because
you can compare different filters all at once. With interactive versions, you
have only one plot, and it's tough to make sense of what you're looking at
when you have a bunch of highlighting all in one place.

P Download
GGobi for free at
http://ggobi.org .

You know the first step. Before doing any visualizing, you need data. Load
our education data into R with read.csv() .
education <- read.csv(“http://datasets.flowingdata.com/education.csv”,
header=TRUE)
education[1:10,]

There are seven columns. The first is for state name, including “United
States” for the national average. The next three are average reading,
mathematics, and writing SAT scores. The fifth column is percentage
of graduates who actually take the SAT, and the last two columns are
pupil-to-staff ratio and high school dropout rate. What you're interested
in is whether any of these variables are related and if there are any clear
groupings. For example, do states with high dropout rates tend to have low
SAT scores on average?

The base distribution of R doesn't supply a straightforward way to use par-
allel coordinates, but the lattice package does, so use that. Go ahead and
load the package (or install if you haven't already).
library(lattice)

Great, now this will be super easy. The lattice package provides a
parallel() function that you can quickly use.

parallel(education)