R Data Visualization Cookbook

We can start by powering up our R studio, which is an Integrated Development Environment (IDE) for R. If you have not downloaded Rstudio, then I would highly recommend going to http://www.rstudio.com/ and downloading it.

To install a package in R, we will use the install.packages() function. Once we install a package, we will have to load the package in our active R session; if not, we will get an error. The library() function allows us to load the package in R.

The install.packages() function comes with some additional arguments but, for the purpose of this book, we will only use the first argument, that is, the name of the package. We can also load multiple packages by using install.packages(c("plotrix", "RColorBrewer")). The name of the package is the only argument we will use in the library() function. Note that you can only load one package at a time with the library() function unlike the install.packages() function.

It is hard to remember all the functions and their arguments in R, unless we use them all the time, and we are bound to get errors and warning messages. The best way to learn R is to use the active R community and the help manual available in R.

To understand any function in R or to learn about the various arguments, we can type ?<name of the function>. For example, I can learn about all the arguments related to the plot() function by simply typing ?plot or ?plot() in the R console window. You will now view the help page on the right side of the screen. We can also learn more about the behavior of the function using some of the examples at the bottom of the help page.

If we are still unable to understand the function or its use and implementation, we could go to Google and type the question or use the Stack Overflow website. I am always able to resolve my errors by searching on the Internet. Remember, every problem has a solution, and the possibilities with R are endless.

In order to generate a numeric vector in R, we can use the C() notation to specify it as follows:

x = c(1:5) # Numeric Vector

To generate a character vector, we can specify the same within quotes (" ") as follows:

y ="I am Home" # Character Vector

To generate a complex vector, we can use the i notation as follows:

c = c(1+3i) #complex vector

A list is a combination of a character and a numeric vector and can be specified using the list() notation:

z = list(c(1:5),"I am Home") # List

The missing values are represented in R by NA. When we download data, it may have missing data and this is represented in R by NA:

z = c( 1,2,3, NA,5,NA) # NA in R is missing Data

To detect missing values, we can use the install.packages() function or is.na(), as shown:

complete.cases(z) # function to detect NA
is.na(z) # function to detect NA

To remove the NA values from our data, we can type the following in our active R session console window:

clean <- complete.cases(z)
z[clean] # used to remove NA from data

Please note the use of square brackets ([ ]) instead of parentheses.

In R, not a number is abbreviated as NaN. The following lines will generate NaN values:

##NaN
0/0
m <- c(2/3,3/3,0/0)
m

The is.finite, is.infinite, or is.nan functions will generate logical values (TRUE or FALSE).

is.finite(m)
is.infinite(m)
is.nan(m)

The following line will generate inf as a special value in R:

## infinite
k = 1/0

complete.cases(z) is a logical vector indicating complete cases that have no missing value (NA). On the other hand, is.na(z) indicates which elements are missing. In both cases, the argument is our data, a vector, or a matrix.

R also allows its users to check if any element in a matrix or a vector is NA by using the anyNA() function. We can coerce or assign NA to any element of a vector using the square brackets ([ ]). The [3] input instructs R to assign NA to the third element of the dk vector.

A vector in R is defined using the c() notation as follows:

vec = c(1:10)

A vector is a one-dimensional array. A matrix is a multidimensional array. We can define a matrix in R using the matrix() function. Alternatively, we can also coerce a set of values to be a matrix using the as.matrix() function:

mat = matrix(c(1,2,3,4,5,6,7,8,9,10),nrow = 2, ncol = 5)
mat

To generate a transpose of a matrix, we can use the t() function:

t(mat) # transpose a matrix

In R, we can also generate an identity matrix using the diag() function:

d = diag(3) # generate an identity matrix

We can nest the rep () function within matrix() to generate a matrix with all zeroes as follows:

zro = matrix(rep(0,6),ncol = 2,nrow = 3 )# generate a matrix of Zeros
zro

We can define our data in the matrix () function by specifying our data as its first argument. The nrow and ncol arguments are used to specify the number of rows and column in a matrix. The matrix function in R comes with other useful arguments and can be studied by typing ?matrix in the R command window.

The rep() function nested in the matrix() function is used to repeat a particular value or character string a certain number of times.

The diag() function can be used to generate an identity matrix as well as extract the diagonal elements of a matrix. More uses of the diag() function can be explored by typing ?diag in the R console window.

The code file provides a lot more functions that can used along with matrices—for example, functions related to finding a determinant or inverse of a matrix and matrix multiplication.

In order to extract any element of a matrix, we can specify the position of that element in R using square brackets. For example, mat[2,3] will extract the element under the second row and the third column. The first numeric value corresponds to the row and the second numeric value corresponds to a column [row, column].

Similarly, to replace an element, we can type the following lines in R:

mat[2,3] = 16

To select all the elements of the second row, we can use mat[2, ]. If we do not specify any numeric value for a column, R will automatically assume all columns.

A data frame in R is a collection of variables. A simple way to construct a data frame is using the data.frame() function in R:

data = data.frame(x = c(1:4), y = c("tom","jerry","luke","brian"))
data

Many times, we will encounter plotting functions that require data to be in a data frame. In order to coerce our data into a data frame, we can use the data.frame() function. In the following example, we create a matrix and convert it into a data frame:

mat = matrix(c(1:10), nrow = 2, ncol = 5)
data.frame(mat)

The data.frame() function comes with various arguments and can be explored by typing ?data.frame in the R console window. The code file under the title Data Frames – 2 provides additional functions that can help in understanding the underlying structure of our data. We can always get additional help by using the R documentation.

We can add or extract any column of a data frame using the dollar ($) symbol, as depicted in the following code:

data = data.frame(x = c(1:4), y = c("tom","jerry","luke","brian"))
data$age = c(2,2,3,5)
data

In the preceding example, we have added a new column called age using the $ operator. Alternatively, we can also add columns and rows using the rbind() and cbind() functions in R as follows:

age = c(2,2,3,5)
data = cbind(data, age)

The cbind and rbind functions can also be used to add columns or rows to an existing matrix.

To remove a column or a row from a matrix or data frame, we can simply use the negative sign before the column or row to be deleted, as follows:

data = data[,-2]

The data[,-2] line will delete the second column from our data.

To re-order the columns of a data frame, we can type the following lines in the R command window:

data = data.frame(x = c(1:4), y = c("tom","jerry","luke","brian"))
data = data[c(2,1)]# will reorder the columns
data

To view the column names of a data frame, we can use the names() function:

names(data)

To rename our column names, we can use the colnames() function:

colnames(data) = c("Number","Names")

In order to import a CSV file in R, we can use the read.csv() function:

test = read.csv("raw.csv", sep = ",", header = TRUE)

Alternatively, read.table() function allows us to import data with different separators and formats. Following are some of the methods used to import data in R:

The first argument in the read.csv() function is the filename, followed by the separator used in the file. The header = TRUE argument is used to instruct R that the file contains headers. Please note that R will search for this file in its current directory. We have to specify the directory containing the file using the setwd() function. Alternatively, we can navigate and set our working directory by navigating to Sessions | Set working directory | Choose directory.

The first argument in the read.table() function is the filename that contains the data, the second argument states that the data contains the header, and the third argument is related to the separator. If our data consists of a semi colon (;), a tab delimited, or the @ symbol as a separator, we can specify this under the sep ="" argument. Note that, to specify a separator as a tab delimited, users would have to substitute sep = "," with sep ="\t" in the read.table() function.

One of the other useful arguments is the row.names argument. If we omit row.names, R will use the column serial numbers as row.names. We can assign row.names for our data by specifying it as row.names = c("Name").

To export data from R, we can use the write.table() function. Please note that R will export the data to our current directory or the folder we have assigned using the setwd() function:

write.table(data, "mydata.csv", sep=",")

The first argument in the write.table() function is the data in R that we would like to export. The second argument is the name of the file. We can export data in the .xls or .txt format, simply by replacing the mydata.csv file extension with mydata.txt or mydata.xls in the write.table() function.

In order to write a simple function in R, we must first open a new R script by navigating to File | New file.

We write a very simple function that accepts two values and adds them together. Copy and paste the code in the new blank R script:

add = function (x,y){
  x+y
}

A function in R should be defined by function(). Once we define our function, we need to save it as a .r file. Note that the name of the file should be the same as the function; hence we save our function with name add.r.

In order to use the add() function in the R command window, we need to source the file by using the source() function as follows:

source('<your path>/add.R')

Now, we can type add(2,15) in the R command window. You get 17 printed as an output.

The function itself takes two arguments in our recipe but, in reality, it can take many arguments. Anything defined inside curly braces gets executed when we call add(). In our case, we request the user to input two variables, and the output is a simple sum.

The logic for if else statements is very simple and is as follows:

if(x>3){
  print("greater value")
}else {
  print("lesser value")
}

We can copy and paste the preceding statement in the R console or write a function that makes use of the if else logic.

The logic behind if else statements is very simple. The following lines clearly state the logic:

if(condition){
#perform some action
}else {
  #perform some other action
}

The preceding code will check whether x is greater than or less than 3, and simply print it. In order to get the value, we type the following in the R command window:

x = 2

The following lines of code multiply each element of x and y and store them as a vector z:

x = c(1:10)
y = c(1:10)
for(i in 1:10){
z[i] = x[i]*y[i]
}

In the preceding code, a calculation is executed 10 times. R performs any calculation specified within {}. We are instructing R to multiply each element of x (using the x[i] notation) by each element in y and store the result in z.

We first define a matrix in R using the following matrix() function:

mat= matrix(1:25, 5,5)

Now, we use the following code to output only those elements where I = J:

for (i in 1:5){
  for (j in 1:5){
    if (i ==j){
      print(mat[i,j])
    }
   }
}

The if statement is nested inside two for loop statements. As we have a matrix, we have to use two for loops instead of just one. The output of the matrix would be values such as 1, 7, 13, and 19.

The apply() function can be used as follows:

mat= matrix(1:25, 5,5)
apply(mat,1,sd)

The lapply() function can be used in the following way:

j = list(x = 1:4, b = rnorm(100,1,2))
lapply(j,mean)

The tapply() function is useful when we have broken a vector into factors, groups, or categories:

tapply(mtcars$mpg,mtcars$gear,mean)

The first argument in the apply() function is the data. The second argument takes two values: 1 and 2; if we state 1, R will perform a row-wise computation; if we mention 2, R will perform a column-wise computation. The third argument is the function. We would like to calculate the standard deviation of each row in R; hence we use the sd function as the third argument. Note that we can define our own function and replace it with the sd function.

With regard to the lapply() function, we have defined J as a list and would like to calculate the mean. The first argument in the lapply() function is the data and the second argument is the function used to process the data.

The first argument in the tapply() function is the data; in our case it is mpg. The second argument is the factor or the grouping; in this case it would be gears. The last argument is the function used to process the data. We would like to calculate the mean of mpg for each unique gear (3, 4, and 5 gears) in the mtcars data.

In the following code, I have highlighted some commonly used parameters:

x=c(1:10)
y=c(1:10)
par(bg = "#646989", las = 1, col.lab = "black", col.axis = "white",bty = "n",cex.axis = 0.9,cex.lab= 1.5)
plot(x,y, pch = 20, xlab = "fake x data", ylab = "fake y data")

Under the par() function, we have set the background color using the bg = argument. The las = argument changes the orientation of the labels. The col.lab and col.axis arguments are used to specify the color of the labels as well as the axis. The cex argument is used to specify the size of the labels and axis. The bty argument is used to specify the box style in R.

To save a plot in the .png format, we can use the png() function as follows:

png("TEST.png", width = 300, height = 600)
plot(x,y, xlab = "x axis", ylab = "y axis", cex.lab = 3,col.lab = "red", main = "some data", cex.main=1.5, col.main = "red")
dev.off()

We have used the png() function to save the plot as a PNG. To save a plot as a PDF, SVG, or JPEG, we can use the pdf(), svg(), or jpeg() functions, respectively.

The first argument in the png() function is the name of the file with the extension, followed by the width and height of the plot. We can now use the plot() function to generate a plot; any subsequent plots will also be saved with a .png extension, unless the dev.off() function is passed. The dev.off() function instructs R that we do not need to save the plots.