Getting It Together

Lab 2G

Directions: Follow along with the slides, completing the questions in blue on your computer, and answering the questions in red in your journal.

Space, Click, Right Arrow or swipe left to move to the next slide.

Putting data together

In the labs so far, we’ve only ever looked at individual data files.
- But often times, we gain additional insights by including additional information from a separate dataset.
In this lab, we will learn how to merge information from our personality color data with our stress/chill data.
Export, upload, import your Personality Color dataset and name it colors.
Then, export, upload, import your Stress/Chill dataset and name it stress.

Looking at Stress/Chill

We would like to analyze the research question:

How do people’s personality colors and/or sports participation affect their stress levels?

We already have data about personality color and a separate dataset about stress.
- What we don’t have is a single dataset with information from both … yet.
We’ll start then by strategizing how to merge our data together.

Deciding how to merge

Before we merge data, we need to decide how we plan to merge it:
We can stack our datasets, that is, take one dataset’s rows and add them to the bottom of the other dataset.
We can also join our datasets horizontally. This is where we take one dataset’s columns and add them to the end of the other dataset’s columns based on matching an ID variable.
- The ID variable will have entries that we use to match observations in both datasets.
To answer the statistical question of interest, would it make more sense to stack or join our colors and stress data?

Finding variables in common:

Look at the names of the variables in each dataset.
- To merge different datasets together, we need to find variables they have in common.
Which variables do the datasets have in common?
Which variable would make sense to merge the datasets together with? Why not the others?

Caution required

Whether stacking or joining, we need to be careful when we merge data:
When stacking data, we need to be absolutely certain that the variables we’re stacking represent the exact same measurements.
- We wouldn’t want to stack height in meters and height in inches, for instance (without converting one to the other).
When joining data, we need to make sure that the id variable in our primary dataset matches to one and only one observation in the joining data.
- Otherwise, R won’t know which observation to match to.

Getting ready

Our goal is to add the variables from the colors data onto the stress data.
Start by ensuring that every user.id in the colors data is unique.
- If there’s a duplicate, have your teacher remove the duplicate from your class’ Web Response Manager and then re-export, upload, import your colors data.
After we add the data from colors to stress, how many rows should our merged data have? Write this number down.

Putting them together

We can use the merge function to join our datasets together using the variables that appear in both sets.
Fill in the blanks below to join the information from the colors data onto the stress data.

merge(____, ____, by = "____")

Assign this merged dataset the name stress_colors.
- Make sure your data has the same number of observations that you wrote down on the previous slide.

Saving your data:

View your merged data and make sure nothing appears to be blatantly wrong with it.
Why didn’t we stack the rows of data instead?
What happens if you swap the order of the datasets in the merge function?
Fill in the blank below to save our stress_colors data for later use.

save(stress_colors, file = "stress_colors.Rda")

Be sure to look in the Files tab to make sure your data was saved.

Moving on

In the next practicum, we’ll begin analyzing our merged data. In the meantime:
Make a few plots using variables from the stress data and facet or group the plots based on variables from the colors data.
Write down the most interesting discovery you make by just exploring your data. Write out how you found your discovery and interpret what it means for the people in your class.
With our colors data, we could answer questions about the typical color scores in your class. Why can we no longer answer this question in our stress_colors data?