Data with Clarity

• Previously, we’ve looked at graphs of entire variables (by looking at all of their values).
  • Doing this is helpful to get a big picture idea of our data.
• In this lab, we’ll learn how to zoom in on our data by learning how to subset.
  • We’ll also learn a few ways to manipulate the plots we’ve been making to make them easier to use for analyses.
• Import the data from your class’ Food Habits campaign and name it food.

Another plotting function

• A dotPlot is another plot that can be used to analyze a numerical variable.
  • Dotplots are better suited for smaller datasets. If datasets are too large, the dots become too small to see.
  • Similarly, distributions with a large spread might impact the readability of the plot.
• Use the dotPlot() function to create a dotPlot of the amount of sugar in our food data.
  • The code to create a dotPlot is exactly like you’d use to make a histogram.
  • Make sure to use a capital P in dotPlot.

More options

• While a dotPlot should conserve the exact value of each data point, sometimes it behaves like a histogram in that it lumps values together.

• Create a more accurate dotPlot by including the nint option.

  • Set nint equal to the maximum value for sugar minus the minimum value for sugar plus one.
    • On your food data spreadsheet, click on the sugar header to sort in ascending order (to obtain minimum).
    • Click on the sugar header again to sort in descending order (to obtain maximum).
  • Use your History pane to see how we included the option nint with the histogram function.

• Pro tip: If the dotPlot comes out looking wonky, try changing the value of the character expansion option, cex.

  • The default value is 1. Try a few values between 0 and 1 and a few more values larger than 1.

Splitting datasets

• In lab 1B, we learned that we can facet (or split) our data based on a categorical variable.

• Split the dotPlot displaying the distribution of grams of sugar in two, by faceting on our observations’ salty_sweet variable.

  • Describe how R decides which observations go into the left or right plot.
  • What does each dot in the plot represent?

Altering the layout

• It would be much easier to compare the sugar levels of salty and sweet snacks if the dotPlots were stacked on top of one another.
• We can change the layout of our separated plots by including the layout option in our dotPlot function.
  • Add the following option to the code you used to create the dotPlot split by salty_sweet.

layout = c(1,2)

• Hint: Use a similar syntax used with the nint option to add the layout option to the dotPlot function.

Subsetting

• Subsetting is a term we use to describe the process of looking at only the data that conforms to some set of rules:
  • Geologists may subset earthquake data by looking at only large earthquakes.
  • Stock market traders may subset their trading data by looking only at the previous day’s trades.
• There’s many ways to subset data using RStudio, we’ll focus on learning the most common methods.

The filter function

• Creating a plot for Salty and a plot for Sweet is useful for comparing Salty and Sweet. What if we want to examine one group by itself?
• The line of code below creates a subset of the food dataset containing only Salty snacks. We will break it down piece by piece in the next few slides.
• Run the line of code below.

food_salty <- filter(food, salty_sweet == "Salty")

• View food_salty and write down the number of observations in it.

So what’s really going on?

• Coding in R is really just about supplying directions in a way that R understands.
  • We’ll start by focusing on everything to the right of the <- symbol

food_salty <- filter(food, salty_sweet == "Salty")

• filter() tells R that we’re going to look at only the values in our data that follow a rule.
• The first blank should be the data we’re going to filter down into a smaller set (based on our rule).
• salty_sweet == "Salty" is the rule to follow.

3 parts of defining rules

• We can decompose our rule, salty_sweet == "Salty", into 3 parts:
  1. salty_sweet, is the particular variable we want to use to select our subset.
  2. "Salty" is the value of the variable that we want to select. We only want to see data with the value "Salty" for the variable salty_sweet.
  3. == describes how we want to relate our variable (salty_sweet) to our value ("Salty"). In this case, we want values of salty_sweet that are exactly equal to "Salty".
• Notice: Values (that are also words) have quotation marks around them. Variables do not.

More on ==

• We can use the head() function to help us see what’s happening when we write salty_sweet == "Salty".
  • head() returns the values of the first 6 observations.
  • The tail() function returns the last 6 observations.
• Run the following code and answer the question below:

head(~salty_sweet == "Salty", data = food)

• What do the values TRUE and FALSE tell us about how our rule applies to the first six snacks in our data? Which of the first six observations were Salty?

Saving values

• To use our subset data we need to save it first.
  • When we save something in R what we are really doing is giving a value, or set of values, a specific name for us to use later.
• The arrow <- is called the “assignment” operator. It assigns names (on the left) to values (on the right).
  • We now focus on everything to the left of, and including, the “<-” symbol.

food_salty <- filter(food, salty_sweet == "Salty")

Saving our subset

food_salty <- filter(food, salty_sweet == "Salty")

• This code then:
  • takes our subset data, (everything to the right of <-) …
  • and assigns the subset data, by using the arrow <- …
  • the name food_salty.
• We can now use food_salty to do anything we could do with the regular food data …
  • but only including those snacks who reported being Salty.
• As a result of assigning the subset data to food_salty, food_salty now appears in the Environment pane. Whenever data is assigned to a variable name, that variable name will appear in the Environment pane.
• Use food_salty to make a dotPlot of the sodium in our Salty snacks.

Including more filters

• We often want to filter our data based on multiple rules.
  • For instance, we might want to filter our food data based on the food being salty AND having less than 200 calories.
• We can include multiple filters to our subsets by separating each rule with a comma like so:

my_sub <- filter(food , salty_sweet == "Salty", calories < 200)

• View the my_sub data we filtered in the above line of code and verify that it only includes salty snacks that have less than 200 calories.

Put it all together

• Use an appropriate dotPlot to answer each of the following questions:

  • About how much sugar does the typical sweet snack have?
  • How does the typical amount of sugar compare when healthy_level < 3 and when healthy_level > 3?

• Because you are now working with subsets of data, it is important to label our plots and make this distinction.

  • We can use the main option to add a title to our plots.
    • Add the following option to the code you used to create the dotPlot of the sugar in Sweet snacks.

main = "Distribution of sugar in sweet snacks"