R Part 3

The total length of the videos in this section is approximately 25 minutes, but you will also spend time running code while completing this section.

You can also view most of the videos in this section at the YouTube playlist linked here. (The barplot video is not on youtube at this point.)

Please download the code file, which is used in the first videos:

Multiple plots

Question 1: How do you know whether you should use boxplot(vec1, vec2, vec3) or boxplot(combinedvec~groups) to plot 3 boxplots on one set of axes?

Output as an object

Question 2: How can you determine what is able to accessed in a saved output using the " $ " syntax?

• Type words that you think should be reasonable, and see if you get an error or not.

• Use the names() command, with your object name in the parentheses

Helpful commands

Question 3: Which of the following outputs would be obtained by the code: 

ifelse(c(1,2,3,4) < 3, "Y", "N")

• "Y" "Y" "N" "N"

• "N" "N" "Y" "Y"

Saving, loading, writing ,reading

Question 4: For which of the following commands should the file name specified end in a '.R' extension? Check all that apply.

• save()

• write()

• load()

• read()

Here is one more idea that will be helpful as you move on. When you change a variable or make a new version of it, the best practice is to create is as a new column in your data set:

dataset$LogIncome<-log(dataset$income)

is much better than

LogIncome<-log(dataset$income)

The latter will not become a new column in your data set, and so it won't be included if you take a subset of your entire data set, reorder the rows, etc.

Next, please download this code file, while will be used in the last video:

Question 5: What does the barplot command want as input, instead of the data itself?

Here is a frequently asked question: How do I reorder the bars in a barplot?

First, note that you can make a barplot by first creating a table, and then using the table as an argument for barplot. For example, if you have a categorical variable about voting and a categorical variable about political view, you might investigate the relationship between these two variables via

mytable<-table(vote, view)

barplot(mytable, legend=TRUE)

By default, R will sort the categories alphabetically. This usually isn't what you want, though. The simplest solution is to reorder the rows and/or columns in the table. For example, if the categories for political view are currently ordered alphabetically: "Extremely Conservative", "Extremely Liberal","Moderately Conservative", "Moderately Liberal","Neutral"

You can reorder as follows:

mytable<-table(vote, view)

mytable<-mytable[ , c(2,4,5,3,1)] # because we want the second column from the original table, and then the fourth, and then the fifth, etc.

barplot(mytable, legend=TRUE)

You can also specify an order for the categories as part of the variable itself. This might be useful because then the categories will appear in the correct order for the rest of your analyses. The two arguments to the function factor are the name of the variable and a vector of the levels in whatever order you prefer.

view<-factor(view, levels=c("Extremely Liberal","Moderately Liberal","Neutral","Moderately Conservative","Extremely Conservative"))

This link expands on how to deal with categories and their order.

During this tutorial you learned:

• To create a set of boxplots visualizing several vectors and change the labels with names= argument

• Review of the $ symbol and how to use it for objects

• About several helpful functions for data manipulation 

• How to import and export csv files

• How to save and load data and R objects as .R files

• How to save graphics as a PDF or jpeg file

• How to make a barplot

Functions in review: 

boxplot(), groups(), by(), apply(), which(), if() { }, ifelse(), save(), load(), write.csv(), pdf(), dev.off(), barplot()