# Your TA will show you how to launch R.
# For most of the lab sessions you will type things in on your own. This is
# painful, but important. Later, you'll be able to cut/paste bits of R code 
# that you will be given. Even when you type things in an R window, note that 
# the up/down arrow keys scroll thru all the commands you have typed in already. 
# So, you don't have to type-in things that you've already typed-in. 

# First, here are some R basics:

(57/276)^2+4.3*.001          # = 0.04695123 . Basic algebraic operations.

## Lines starting with "#" are comments.

## Create variables and assign them value:

## Variable names, function names, etc., should contain only alphanumeric 
## characters (A-Z, a-z, 0-9) and the period ".". Take note, a name cannot 
## start with a digit and ".". R is case-sensitive, i.e., "a" and "A" are
## different.  Here are some more basic operations:

x = 3         ## assignment can be done using either "=" or "<-"
x
y = 1 + 1
y
log10(y)
log2(y)
sqrt(y)

# Variables can be either number, vector, matrix, dataframe, character, or 
# logical expression. No need for their definitions here, for we will come 
# across them gradually.

x = 1:5                   # A few ways of entering data in R:
x
y = c (34, 30, 41, 35, 21)   
y

mean(y)    # sample mean of y -- measures "location."
median(y)  # sample median of y -- another measure of "location."
sd(y)      # sample standard deviation of y -- measures "spread."
           # We may not have done sd in class, yet, but don't worry.
           
                         # For the following, use your up-arrow key.
range(y)                 # Gives two numbers, min and max. 
range(y)[1]              # First component of range(y), i.e., min
range(y)[2]              # Second ..., i.e., max.
min(y)
max(y)

sort(y)

###################################################################

## To get help and examples on any function, say median:

?median           # Explains and usually gives examples at end.
                  # q will get you out of the help page.
                  
help.search("histogram")  # searches all of R (on your computer). q quits.

###################################################################

# R comes with a bunch of data sets. This is how you see what they are:

data()              # Now you know where to get data, if you want some!

# This is how you read-in the R data sets:

data()              # lists all R data sets; q quits
data(USArrests)     # This loads a specific R dataset.  
?USArrests          # Gives info on data set; q quits.
USArrests           # Simply prints all the data onto the screen.
names(USArrests)    # Tells you the name of the variables.         Use up-arrow.
dim(USArrests)      # Tells the (row X column) dimensions.         Use up-arrow.  
USArrests$Murder    # $ selects a given column/variable, by name.  Use up-arrow.
USArrests[,1]       # Same as above, but selects by column number. Use up-arrow.

murder = USArrests$Murder # Selects (by name) and assigns to some variable.
murder                    # shows murder.
x = USArrests[,1] # Selects (by column number) and assigns to some variable.
x                 # Shows arrest.
mean(x)           # Use up-arrow.
sd(x)             # standard deviation of arrest.

####################################################################

# Many of the hw problems will require reading in data from the CD. 
# You can either enter the data via one of the means above, e.g., x= c(1,2,3),
# or read-in data from CD of your textbook, like this:

  dat = read.table(file.choose(),header=T)

# For reading excel files, save the file as .csv, and then read it as

  dat = read.csv(file.choose(),header=F)

# The following shows how you read-in data from a text file. 
# (At the bottom of this file, below, you can see how to change directory.)

dat = read.table("http://www.stat.washington.edu/marzban/390/hist_dat.txt", header=F)
dat

## Now, let's do some histograms on this data.

# A histogram is simply a count (or frequency) plot of different values of a 
# variable. It answers the questions, for example, "How many x=1's occur in the 
# data?", "How many x=2's occur in the data?" Etc. That's easy when the variable 
# (x) is discrete. But what if it's continuous? Then, the answer to all of the
# above questions is "zero." The only reasonable questions are of the type 
# "How many cases fall between 0 and 0.1? How many between 0.1 and 0.2?" Etc.
# The 0.1 is called the class interval, or the bin size. What you will see 
# here is the effect of this bin size on the histogram. In R, instead of 
# specifying the bin size, it's more convenient to simply specify the 
# number of bins, specified by "breaks=". Here are some examples.

x = dat[,1]      # Selects the first column/variable in the file.
x

# Let's turn the "knob" on binsize and see what happens:

                    # For the following, USE YOUR UP-ARROW.
par(mfrow=c(3,3))   # This puts the 9 following histograms on 1 page.
hist(x, breaks = 2) # looks useless
hist(x, breaks = 3)
hist(x, breaks = 4) # looks unimodal and bell-shaped.
hist(x, breaks = 5)
hist(x, breaks = 10) # looks bimodal.
hist(x, breaks = 20)
hist(x, breaks = 30)
hist(x, breaks = 100) # looks bimodal + outlier
hist(x, breaks = 10000) # looks useless

# The above suggests that the data/variable x is probably made-up of
# two different groups. For example, x could be "height", in which case
# the two "humps" (at around breaks = 30 or 100 may be identified as the 
# heights of boys and girls, respectively. This type of analysis is a
# simple form of data-mining, i.e. trying to figure out what's in the data.

##################################################################

# Finally, how to make a "density scale" histogram:

dat = read.table("http://www.stat.washington.edu/marzban/390/hist_dat.txt", header=F)
x = dat[,1]

# Recall that a density scale histogram is a relative frequency histogram where
# the y-axis is also divided by the bin size. R does it like this:

   par(mfrow=c(1,2))
   hist(x)               # "regular" histogram
   hist(x,freq=FALSE)    # density scale histogram

# You can see that the shape is the same, but the advantage of the density
# is that the area under it is 1.
# By the way, since by default R takes the binsizes to be constant, the
# above density histogram is also a relative frequency histogram.

###############################################################

# Take qz1. 

q()           # This will get you out of R altogether.  

####################################################################
####################################################################

# Time permitting:

is.vector(x)       # Checking to see if something is a vector.

                   # To make a sequence of numbers, use c() or seq():
y = seq (from = 0, to = 1, by = 0.1)  # Or simply,  y = seq (0, 1, 0.1)

                   # If you want the breaks in specfic places:
hist(murder, breaks=seq(0, 20, by=3)) 
hist(murder, breaks=c(0,1,5,10,20))

# This is how you read-in data from a file (e.g. hist_dat.txt) in some
# directory:

setwd("G://Documents/2009winter/STAT 390/Lab")  # This is an example!!!
dat = read.table("hist_dat.txt", header=F)

###########################################################################