# Plot and Test Samples in R

README.md

Plot and Test Samples in R

Install R

https://cran.r-project.org/bin/

Setup

chmod a+x plot.R

Run

./plot.R <samples.csv
open boxplot.svg
open histogram.svg

Notes:

The wilcox.test does not assume the data to be normal and is only answering a yes/no for whether the sample sets came from the same overall set (the null hypothesis).

The dots in the box plot are outliers defined as outside 1.5 x IQR (Interquartile Range; the length between the 25th and 75th percentile).

boxplot.svg

histogram.svg

plot.R

#!/usr/bin/env Rscript
samples = read.csv('stdin')

# by default levels for string factor is alphabetical
# but we want experiment to be x
samples$set <- relevel(samples$set, ref = "experiment")

# test whether the true location of the distribution is shifted
# does not require data to be normally distributed
wilcox.test(µs ~ set, data=samples, conf.int=TRUE)

svg(file = "boxplot.svg")
boxplot(µs ~ set, data=samples)
dev.off()

require(lattice)
svg(file = "histogram.svg")
histogram(~µs | set, data=samples, breaks=nclass.FD)
dev.off()

samples.csv