Getting Started

Create a new markdown document to answer questions from this lab, and save it in a sensible place on your computer. If you select the “From Template” option there is a “Lab Report” template that will work very nicely for you!

In this lab you will use the ggplot2 package to generate graphics. The “The Grammar of Graphics,” is the theoretical basis for the ggplot2 package. Much like how we construct sentences in any language by using a linguistic grammar (nouns, verbs, etc.), the grammar of graphics allows us to specify the components of a statistical graphic.

In short, the grammar tells us that:

A statistical graphic is a mapping of data variables to aesthetic attributes of geometric objects.

We can break a graphic into the following three essential components:

  • data: the data-set comprised of variables that we plot
  • geom: this refers to our type of geometric objects we see in our plot (points, lines, bars, etc.)
  • aes: aesthetic attributes of the geometric object that we can perceive on a graphic. For example, x/y position, color, shape, and size. Each assigned aesthetic attribute can be mapped to a variable in our data-set.

Remember: Unless I indicate “text only”, you are expected to include all code you need to answer questions. In other words, if you are asked a question about a graphic, include the code for that graphic.

Resources: You will probably want to refer to the R for Data Science data visualization chapter.

R packages

We will use the following packages, which are all contained in the tidyverse package. The tidyverse package is actually a collection of packages that share an underlying common philosophy of coding and “tidy” data. Today we are using the following:

  • dplyr: for data wrangling
  • ggplot2: for data visualization
  • readr: for reading in data

Remember to use the library command to load tidyverse, openintro packages into your environment.

The data

Today we will practice data visualization using data on births from the state of North Carolina. Copy, paste and run the code below to load the data into your workspace (i.e. using the console). These data are found in the openintro package.

data(ncbirths)

The data set that shows up in your Environment is a large data frame. Each observation or case is a birth of a single child.

The data command instructs R to load some data built into a package. The workspace environment in the upper right hand corner of the R Studio window should now list a data set called ncbirths with 1000 observations (rows or cases) and 13 variables (columns).

How to look at data in R

Take a glimpse

You can see the dimensions of this data frame (# of rows and columns), the names of the variables, the variable types, and the first few observations using the glimpse function. There are LOTS of ways to get a sense of your data, glimpse is just one of them. You’ll pick up more, and identify your favorites, as you get more practice.

glimpse(ncbirths)

We can see that there are 1000 observations and 13 variables in this data set. The variable names are fage, mage, mature, etc. This output also tells us that some variables are numbers…some specifically integers <int>, others are numbers with decimals <dbl>. Some of the variables are factors <fct> (categories). It is a good practice to see if R is treating variables as factors <fct>; as numbers <int> or <dbl> (basically numbers with decimals); or as characters (i.e. text) <chr>. We can change these types if we don’t like them.

  1. What type of variable is R considering the variable habit to be? What variable type is visits? (answer with text only, but use the glimpse function in the console to get the answer)

The data viewer

You can view the data by clicking on the name ncbirths in the Environment pane (upper right window). This will bring up an alternative display of the data set in the Data Viewer (upper left window). R has stored these data in a kind of spreadsheet called a data frame. Each row represents a different birth: the first entry or column in each row is simply the row number (it’s a different color), the rest are the different variables that were recorded for each birth. You can close the data viewer by clicking on the x in the upper left hand corner.

It is a good idea to try kitting your document from time to time as you go along! Go ahead, and make sure your document is knitting. Note that knitting automatically saves your .Rmd file, too.

Types of graphs

We will explore three different types of graphs initially.

scatterplots
boxplots
histograms

Scatterplots

Scatterplots allow you to investigate the relationship between two numerical variables. While you may already be familiar with this type of plot, let’s view it through the lens of the Grammar of Graphics. Specifically, we will graphically investigate the relationship between the following two numerical variables in the births data frame:

  • weeks: length of a pregnancy on the horizontal “x” axis and
  • weight: birth weight of a baby in pounds on the vertical “y” axis
ggplot(data = ncbirths, aes(x = weeks, y = weight)) + 
  geom_point()

Let’s view this plot through the grammar of graphics. Within the ggplot() function call, we specified:

  • The data frame to be nc by setting data = ncbirths
  • The aesthetic mapping determines the visuals of the plot with aes(x = weeks, y = weight)
  • The variable weeks maps to the x-position aesthetic
  • The variable weight maps to the y-position aesthetic.

We also add a layer to the ggplot() function call using the + sign. The layer in question specifies the geometric object here as points, by specifying geom_point().

Finally, we can also add axis labels and a title to the plot like so. Again we add a new layer, this time a labs or labels layer.

ggplot(data = ncbirths, aes(x = weeks, y = weight)) + 
  geom_point() + 
  labs(x = "Length of pregnancy (in weeks)", y = "Birth weight of baby (lbs)", 
       title = "Relationship between pregnancy duration and newborn weight")
  1. Is there a positive or negative relationship between these variables? (text only to answer)

  2. Make a graph showing weeks again on the x axis and the variable gained on the y axis (the amount of weight a mother gained during pregnancy). Include axis labels with measurement units, and a title.

  3. Study the code below, and the resulting graphical output. Note that I added a new argument of color = premie inside the aesthetic mapping. The variable premie indicates whether a birth was early (premie) or went full term. Please answer with text:

    A. What did adding the argument color = premie accomplish?

    B. How many variables are now displayed on this plot?

    C. What appears to (roughly) be the pregnancy length cutoff for classifying a newborn as a “premie”” versus a “full term”?

ggplot(data = ncbirths, aes(x = weeks, y = gained, color = premie))+ 
  geom_point() + 
  labs(x = "Pregnancy length (wks)", y = "Maternal weight gain (lbs)")
  1. Make a new scatterplot that shows a mothers age on the x axis (variable called mage) and birth weight of newborns on the y axis (weight). Color the points on the plot based on the gender of the resulting baby (variable called gender). Does there appear to be any strong relationship between a mother’s age and the weight of her newborn? Does the sex of the child seem to be a factor?

Histograms

Histograms are useful plots for showing how many elements of a single numerical variable fall in specified bins. This is a very useful way to get a sense of the distribution of your data. Histograms are often one of the first steps in exploring data visually.

For instance, to look at the distribution of pregnancy duration (variable called weeks):

ggplot(data = ncbirths, aes(x = weeks))+ 
  geom_histogram()

A few things to note here:

  • There is only one variable being mapped in aes(): the single numerical variable weeks. You don’t need to compute the y-aesthetic: R calculates it automatically.
  • We set the geometric object as geom_histogram()
  • The warning message encourages us to specify the number of bins on the histogram, as R chose 30 for us.

We can change the binwidth (and thus the number of bins), as well as the colors.

ggplot(data = ncbirths, aes(x = weeks))+ 
  geom_histogram(binwidth = 1, color = "white", fill = "steelblue")

Note that none of these arguments went inside the aesthetic mapping argument as they do not specifically represent mappings of variables to visual properties.

  1. Inspect the histogram of the weeks variable. Answer each of the following with text only.

    A. The y axis is labeled count. What is specifically being counted in this case? Hint: think about what each case is in this data set.

    B. What appears to be roughly the average length of pregnancies in weeks?

    C. If we changed the binwidth to 100, how many bins would there be? Roughly how many cases would be in each bin?

  2. Make a histogram of the birth weight of newborns (which is in lbs), including a title and axis labels.

Faceting

Faceting is used when we’d like to create small multiples of the same plot over a different categorical variable. By default, all of the small multiples will have the same vertical axis.

For example, suppose we were interested in looking at whether pregnancy length varied by the maturity status of a mother (column name mature). This is what is meant by “the distribution of one variable over another variable”: weeks is one variable and mature is the other variable. In order to look at histograms of weeks for older and more mature mothers, we add a plot layer facet_wrap(~ mature, ncol = 1). The ncol = 1 argument just tells R to stack the two histograms into one column.

ggplot(data = ncbirths, aes(x = weeks)) +
  geom_histogram(binwidth = 1, color = "white", fill = "steelblue") +
  facet_wrap(~ mature, ncol = 1)
  1. Make a histogram of newborn birth weight split by gender of the child. Set the binwidth to 0.5. Which gender appears to have a slightly larger average birth weight?

Boxplots

While histograms can help to show the distribution of data, boxplots have much more flexibility, and can provide even more information in a single graph. The y aesthetic is the numeric variable you want to include in the boxplot, and the x aesthetic is a grouping variable. For instance, below we set gender as the aesthetic mapping for x, and gained as the aesthetic mapping for y. This creates a boxplot of the weight gained for mothers that had male and female newborns. Note that the fill argument is not necessary, but sets a color for the boxplots.

ggplot(data = ncbirths, aes(x = gender, y = gained)) +
  geom_boxplot(fill = "sienna")

Take some time to familiarize yourself with the different parts of the boxplot:

  • The bottom of the “box” portion represents the 25th percentile (1st quartile).
  • The horizontal line in the “box” shows the median (50th percentile, 2nd quartile).
  • The top of the “box” represents the 75th percentile (3rd quartile).
  • The height of each “box”, i.e. the value of the 3rd quartile minus the value of the 1st quartile, is called the interquartile range (IQR). It is a measure of spread of the middle 50% of values. IQR is another measure of the spread of the data: longer boxes indicating more variability.
  • The “whiskers” extending out from the bottoms and tops of the boxes represent points less than the 25th percentile and greater than the 75th percentiles respectively. They extend out no more than 1.5 x IQR units away from either end of the boxes. The length of these whiskers show how the data outside the middle 50% of values vary. Longer whiskers indicate more variability.
  • The dots represent values falling outside the whiskers or outliers. The definition of an outlier is somewhat arbitrary and not absolute. In this case, they are defined by the length of the whiskers, which are no more than 1.5 x IQR units long.
  1. Make a boxplot of the weight gained by moms, split by the maturity status of the mothers (mature). Include axis labels and a title on your plot. Is the median weight gain during pregnancy larger for younger or older moms?

  2. Make a boxplot of pregnancy duration in weeks by smoking habit. Is the duration of pregnancy more variable for smokers or non-smokers? (i.e. which group has the greater spread for the variable weeks?). What do you think the “NA” means?


More exercises

For the following, you need to determine which type of plot to use, make the plot, and answer any questions with text. The last few exercises require you to duplicate plots I’ve given using a few other datasets. The ggplot2 cheatsheet is a great guide to choosing the right plot (Help|Cheatsheets|Data visualization with ggplot2). It’s split by the number of variables in your plot as well as the type of variables.


  1. Using a data visualization, visually assess: Is the variable for father’s age (fage) symmetrical, or does it have a skew?

  2. A. Using a data visualization, visually assess: (in this sample) is the median birth weight of babies greater for white or non-white mothers (variable called whitemom)? Why do you think there is an “NA” group? B. Discuss (in your group and in your report) whether you view any ethical dillemmas with the variable whitemom. Some questions to consider: how was this information originally collected? Would findings related to this variable be able to help all women or just some women?

  3. Using a data visualization, visually assess: (in this sample) as a mother’s age (mage) increases, does the duration of pregnancy (weeks) appear to decrease? Hint: Try using geom_jitter to see more of the data. You don’t need to set any arguments for this.

  4. Recreate the plots below…

  • Graph 1: from the mpg dataset in the tidyverse package.
  • Graph 2: from the economics dataset in the tidyverse package. There is info you need in the caption.
  • Graph 3: from the diamonds dataset in the tidyverse package. This one is harder. You need to look up how to make a density plot in ggplot2 (and what it is). There is info you need in the caption.



When you are finished with the lab, go to the very top and change the output from html_document to pdf_document. The pdf document doesn’t look as nice, but it is easier to grade and upload to schoology. Now turn in this PDF file to Schoology. Note the due date and time. If Schoology says it’s late, it’s late. Make sure your final Markdown document Knits properly and shows all your work. Look through it to make sure everything looks organized and professional. Also remember that if you needed output (graphs, numeric output, etc.) to answer a question, the code to generate that output needs to be in the lab report. Other code should not be included.