Chapter 2: Introducton to the research process using R

# Tell R where you will be working.  R will look here for your data:
setwd("~/Documents/Books/Presenting/data/rData/")
# On a machine running Windows, the path would be something like:
# setwd("C:\\Projects\\Books\\Presenting\\data\\rData\\")
 
# Load the dataset rhinitis out of the current folder/directory into R's 
# interactive working memory/environment. If you have not set the working
# directory with the setwd() function, you could type something like this on a
# Mac:
# load("~/Documents/Books/Presenting/data/rData/rhinitis.RData") or
# or this on a Windows machine:
# load("C:\\Projects\\Books\\Presenting\\data\\rData\\rhinitis.RData") 
load("rhinitis.RData")
 
# The head() function shows the beginning of an object.  Here it shows the first
# 14 records of a dataframe.  You also use head() to show the first few values
# in other kinds of objects; for example a vector of numbers or character 
# strings.
head(rhinitis, n = 14)  # Display the first 14 records.

setwd("~/Documents/Books/Presenting/data/rData/")
load("rhinitis.RData")
 
head(rhinitis, n = 14) 

setwd("~/Documents/Books/Presenting/data/rData/")
load("rhinitis.RData")
 
# The code below adds a variable called whenrhinFactor to the rhinitis dataset.
# The categorical factor variable has the values 1, 2 and 3, which are labeled 
# with phrases describing the 3 seasons.
 
# The factor() function is applied to the whenrin variable in the rhinitis 
# dataframe.  factor() uses a comma-delimited set of details. A programmer calls 
# the details that a function understands "arguments". Notice the ( following
# the word factor and the commas ending the lines with the arguments until 
# the closing ) .
rhinitis$whenrhinFactor <- factor(rhinitis$whenrhin,  
                                  # Actual values in the variable are 1, 2, 3.
                                  # The c() function can glue together a set of
                                  # similar things, for example, a bunch of 
                                  # numbers, to make a vector.  The levels 
                                  # argument is expecting a vector.
                                  levels = c(1, 2, 3),      
                                  # Remember that the c() function stores 
                                  # similar things as a vector. In this case,
                                  # the vector is storing character strings for
                                  # the labels argument.
                                  labels = c("Dry season", 
                                             "Wet season", 
                                             "Anytime")
                                  )  # This ) is the end of the factor function.
 
# The comment() function is used to add a descriptive label to the new variable. 
# Sadly, it is not typically automatically shown in output. To later display the 
# comment, type comment(rhinitis$whenrhinFactor).
comment(rhinitis$whenrhinFactor) <- "When get rhinitis"
 
rhinitis$rhinitisFactor <- factor(rhinitis$rhinitis, 
                                  levels = c(0, 1) , 
                                  labels = c("No", "Yes")
                                  )
comment(rhinitis$rhinitisFactor) <- "Rhinitis with a cold in last 12 months"
 
# The code below makes a simple summary table.
 
# The with() function takes two arguments, the name of a dataframe and a 
# function that will be applied to variables in that dataframe.  
 
# The with() function, used below, allows you to work with the variables within 
# a dataframe without having to repeatedly mention the data frame.  The code 
# below could have been written as:
# table(rhinitis$whenrhinFactor,   # rows of the table
#       rhinitis$rhinitisFactor,   # column of the table
#       dnn = c(comment(whenrhinFactor),   # row label
#               comment(rhinitisFactor))   # column label
#       )                                  # closing ) for table function
#
# Another option is to attach the dataframe at the "top" of the search list with
# code like this:
#       attach(nonResponse)
#       table(whenrhinFactor,   # rows of the table
#             rhinitisFactor,   # column of the table
#             dnn = c(comment(whenrhinFactor),   # row label
#                     comment(rhinitisFactor))   # column label
#             )                                # closing ) for table function
#       detach(nonResponse)
# That strategy can work if you are only working with a single dataframe but it 
# can lead to problems. Problems will ensue if you do not detach the data when 
# you have variables with the same name in multiple dataframes.  For example,
# if you are working to predict who has a sexually transmitted infection and 
# if you have a variable called sex in a dataframe called coitus and another 
# variable called sex in a dataframe called demographics, you can easily 
# accidentally use a person's gender when you intended to use an indicator for 
# sexual activity. Look here a good brief introduction:
#     http://www.r-bloggers.com/to-attach-or-not-attach-that-is-the-question/
with(rhinitis,               # name of data frame with the data
     # The table() function does simple cross tabulation tables.
     table(whenrhinFactor,   # rows of the table
           rhinitisFactor,   # column of the table
           dnn = c(comment(whenrhinFactor),   # row label
                   comment(rhinitisFactor))   # column label
           )                                  # closing ) for table function
     )                                        # closing ) for with function  
 
 
 
 
# The code below makes a good looking summary table.
 
# The library() function attaches a package full of functions (and sometimes 
# dataframes) to R's "search list".  When you type CrossTable(), R searches that
# list to figure out what to do.  If you have not attached the package that 
# included a riskratio() function, R will complain that it can not find the 
# function "CrossTable".  
# You can encounter problems when two packages include functions with the same  
# name.  R will give a warning like:
#       The following object(s) are masked ...
# when you attach a package that includes a function that is already in the 
# search list.  R calls this problem masking.  The detach() function, shown  
# below, can be used to remove packages from the search list and help avoid this 
# problem.  The CrossTable() function is unique enough that I do not bother to 
# detach and reattach it repeatedly.  
 
# Load the gmodels package to get the fancy CrossTable() function. 
library(gmodels)
 
with(rhinitis, 
     # The CrossTable() function generates frequency tables and it can do
     # many common cross tabulation statistics.
     CrossTable(whenrhinFactor,    # rows of the table
                rhinitisFactor,    # columns of the table
                prop.r=FALSE,      # don't show row percentages
                prop.c=FALSE,      # don't show column percentages
                prop.t=FALSE,      # don't show total percentages
                # don't show each cell's impact on a chi-square test
                prop.chisq=FALSE,
                dnn = c(comment(whenrhinFactor),   # row label
                        comment(rhinitisFactor)    # column label)
                        )    
                )
     )
 
# Make a dataset (an R dataframe object) that has the heart variables.  
 
# The subset() function takes a dataframe as its first argument and it can use
# an argument called select to select only a subset of variables/columns. It 
# can also include a "logic check" argument called subset which is used to 
# determine which  records/rows to include.  That option appears in later 
# chapters.
hearts <- subset(rhinitis,                          # the dataframe name
                 select = c(syst1, syst2, syst3,    # variables to keep
                            diast1, diast2, diast3, 
                            pulse1, pulse2,  pulse3
                            )
                 )
 
# People who hate typing could type this instead:
#      hearts <- rhinitis[, 1:7]
# The [ , ] after the dataframe name can be used to select which rows and 
# columns are used.  The " [, " indicates that there is
# no filter on the rows.  So, include all subjects.  The ", 1:7]" is a filter on 
# which columns to include.  The analysis will be limited to the 1st through the
# 7th columns.
 
# The summary() function pays attention to the type of object it is being asked
# summarize .  Here, because the object is a data frame, it produces a 5 number 
# summary of the variables. It is missing standard deviation.  So another 
# function needs to be tried (the describe function). In general it is a good
# plan to try the summary() function to describe objects.
summary(hearts)
 
 
# Load psych package to get the describe() function.  This masks many objects in
# other packages including the widely used describe() function whith is in the 
# Hmisc package.
library(psych)
 
# The describe() funcion in the psych package produces a more complete summary,
# including a standard deviation.
describe(hearts)   
 
# Detach the package to prevent masking conflicts.
detach("package:psych", unload=TRUE)  
 
 
 
with(rhinitis, 
     hist(diast2)   # The hist() function draws a simple histogram.
     )
 
# Load the ggplot2 package to get fancy graphics functions including qplot()
# and theme_bw() which are used below. The syntax/grammar used by ggplot2 is 
# somewhat different from the rest of R.  The basic idea with ggplot2 is that 
# you add more and more details to get basic geometric objects to behave.  It, 
# perhaps even more than the rest of R, is evolving and features get devaluated.  
# To learn more about ggplot2 look here: 
# http://ggplot2.org/ 
# http://www.statmethods.net/advgraphs/ggplot2.html
library(ggplot2)  
 
with(rhinitis,      
     qplot(diast2,                      # make a quickplot of the dist2 variable
           geom = "histogram",          # the geometry is a histogram
           main = "ggplot2 histogram",  # the main title
           xlab = "Diastolic 2",        # x-axis label
           ylab = "Frequency",          # y-axis label
           binwidth = 10                # how wide to make the bins/bars
           ) +                          # finish quickplot but add stuff
           theme_bw()                   # change to black on white color scheme   
     )