##########################
##### R Tutorial     #####
##### 2010-01-28     #####
##### PubH 7440      #####
##### Laura Hatfield #####
##########################

#####################
##### Running R #####
#####################
## Comments begin with a hash/pound sign
getwd()
# Shows the current working directory
dir()
# Displays its contents
ls()
# Lists the R objects in the current workspace
source("test.R")
# Runs a code file
ls()
library(MASS)
# Loads a previously installed package
install.packages("BRugs")
# Installs a new package

#### Basic Data Structures ####
## Basic data structures
family <- list()
family$members <- data.frame(
	first=c('Jeff','Martha','Laura','Lydia','Rachel'),
	year=c(1958,1959,1983,1985,1987))
# Creates a two-column data frame
# Each column is named and stores a different type of data
# c() is called the collection function
family[['members'']

family$awesomeness <- rep(1e20,dim(family$members)[1])
# Creates a vector list item
# by repeating the same value
# 1e20 is R's scientific notation
# the number of columns of the family data frame is
# how many times 1e20 will be repeated
family$awesomeness

family$correl <- matrix(c(
	1,.2,.8,.3,.7,
	.2,1,.4,.7,.3,
	.8,.4,1,.1,.6,
	.3,.7,.1,1,.3,
	.7,.3,.6,.3,1),5,5)
# This list item is a matrix
# R fills a matrix with leftmost index moving fastest
# The first argument is a vector of values, 
# second is number of rows, 
# third is number of columns

isSymmetric(family$correl)
# A matrix is a true mathematical object
solve(family$correl)
# Inverts the matrix

rownames(family$correl) <- colnames(family$correl) <- family$members$first
family$correl['Jeff','Martha']
# Matrices may have row and column names
# for convenient extraction

aspirin <- read.table("http://www.biostat.umn.edu/~brad/data/aspirin_data.txt",
	sep="\t",header=F)
# Load a data table from the class website
# Various options allow one to read in a variety of plain text
# data sets
aspirin
class(aspirin)
attributes(aspirin)
# Displays the contents, class, and attributes of the object
names(aspirin) <- c('patient','before','after')
# Assigns strings to the 'names' attribute

#################################
##### Statistical Functions #####
#################################

Z <- rnorm(1000,mean=5,sd=10)
# Draws a random sample of normals
hist(Z,freq=F)
# Plots them in a histogram
dens <- dnorm(0,mean=5,sd=10)
abline(h=dens,col='red')
# Computes and adds lines to denote the value
# of the normal density at a given point
abline(v=0,col='red')
text(-20,dens,paste("f(0)=",round(dens,4)),pos=1,col='red')
# Prints the rouded value of the density and labels it on the plot
lines(seq(-30,40,l=100),dnorm(seq(-30,40,l=100),5,10),col='blue')
# Adds the pdf curve to an existing plot
# Note: seq() creates a list of numbers
# here, from -30 to 40 with length 100
text(20,.03,"f(Z)",pos=1,col='blue')
polygon(x=c(seq(-30,0,l=100),seq(0,-30,l=100)),
  y=c(dnorm(seq(-30,0,l=100),5,10),rep(0,100)),density=10,col='green')
# Draws a shaded area
text(-20,.01,paste("F(0)=",round(pnorm(0,5,10),4)),col='green')
# Prints the (labeled) value of the CDF that was shaded
arrows(qnorm(.975,5,10),.015,qnorm(.975,5,10),0,code=2,col='orange')
text(qnorm(.975,5,10),.015,paste("q(.975)=",
	round(qnorm(.975,5,10),2)),col='orange',pos=4)
# Plots the location of a the 97.5th normal quantile

##############################
##### Simple Programming #####
##############################

myfunction <- function(x){
  for (i in 1:length(x)){
    if (x[i] > 100) { 
      print(paste("Subject ",i,", you have a fever of ",x[i],".",sep=""))
    }
    else {
      print(paste("Go to class, subject ",i,"!",sep=""))
    }
  }
}
# Uses logic and looping to print out information 
# based on an input vector of values
myfunction(aspirin[,2])
myfunction(aspirin[,3])
# Applies the function to the aspirin data before and after

#######################################
##### Simple Statistical Analysis #####
#######################################
reg.results <- lm(aspirin[,2]~1)
reg.results
attributes(reg.results)
summary(reg.results)