This is the website of Yihui Xie - http://yihui.name. I'm currently a first-year PhD student in the Department of Statistics, Iowa State University.
› Current Work
Statistics does not mean whole pages of Greek letters; there are a lot of things to see. I have been working on an R package named animation which aims at demonstrating statistical theories and methods through animations, and turning statistical ideas into animations.
The technical implementation looks simple but effective, and the critical point is how to express the statistical ideas clearly in animations, which makes me think over and over again.
## Brownian Motion follows cum. sums of Gaussian rv's
bm.mat = replicate(2, apply(matrix(rnorm(2000), 200, 10),
2, cumsum), simplify = FALSE)
## plot each step and pause for 0.05 secs
for (i in seq_len(200)) {
plot(bm.mat[[1]][i, ], bm.mat[[2]][i, ],
xlim = range(bm.mat[[1]]),
ylim = range(bm.mat[[2]]),
pch = 19, ann = FALSE, axes = FALSE)
Sys.sleep(0.05)
}
When people
know what I'm working on, they are often surprised: don't all statisticians derive mathematical formulae every day? An answer to this question is:
"I'm not an outlier; I just haven't found my distribution yet!
This semester I begin to work on graphics based on Qt, which seems to be faster than R in terms of graphics. There are a series of packages that can connect R to Qt, and my current job is to rewrite some familiar graphics (e.g. parallel coordinate plot) based on these packages.
› Resources
- The R Project (Statistical Computing & Graphics)
- The Omegahat Project (Infrastructure for Stat Computing)
- Milestones of Stat Graphics (History)
› Contact Me
3211 Snedecor Hall, Department of Statistics, Iowa State University
Ames, 50011, IA
Email: xie@yihui.name; Tel: 515-203-6609