A Happiness Survey

hapsur <- read.table("http://nathanieldphillips.com/wp-content/uploads/2016/04/hapsur.txt")

In this WPA, you’ll analyze data from a survey about happiness. In the survey, 500 people were asked the following questions:

age - How old are you
sex - What is your sex?
exercise - How many hours a week do you exercise on average?
relationship - Are you in a long-term romantic relationship?
drinks - How many alcoholic drinks d you consume a week on average
happiness - On a scale of 0 to 100, how happy are you with your life?

Install and load the latest version of the yarrr package

A. You’ll need to use the yarrr package in this WPA. It has been updated in the past few days so you should reinstall it.

library(devtools)
install_github("ndphillips/yarrr")

B. Now that you’ve installed the latest version you’re good to go. However, as always you need to load the package before you can use it. Load the package with the library() function:

library(yarrr)

Load the data into R

C. Create a new project called wpa.Rproj. You’ll use this project for this and all future wpas.

D. Navigate to the folder containing wpa.Rproj, and create two new folders called data and scripts. You’ll save all future datasets and wpas in these two folders.

E. Now it’s time to download the data for this WPA. You can find the data stored as a tab-delimited text file at http://nathanieldphillips.com/wp-content/uploads/2016/04/hapsur.txt. Download the file and save it to the data folder in your wpa project.

F. Open a new R script and save it the scripts folder under the name WPA5.

G. Using read.table(), load the happiness survey data into R as a new object called hapsur

Histograms

Open the help menu for the histogram function
Create a boring histogram of the exercise data by only specifying the exercise data without any additional arguments.
Now, create a more interesting histogram by making the following changes
- Add labels to the x-axis and the main plot title
- Change the x limits to 0 and 20,
- Change the bar colors (maybe ‘coral’)
- Make the bar border (maybe ‘white’)
- Turn off the y-axis by adding the argument yaxt = ‘n’
Now, create two separate histograms of the exercise data on top of each other. One for men and one for women.
- Create two new dataframes suvey.men and hapsur.women that only contain data from men and women respectively.
- Create a histogram of the age data for men. Set the color of the bars to transparent(“red”, .5) (or use any other color you want)
- Add a new histogram of the age data for women by including the argument add = T. Set the color of the bars to transparent(“blue”, .5) (or use any other color you want)
- Add a legend with the following code
The abline() function allows you to add lines to plots. For example:

Using abline(), add a line showing the mean of each of the two histograms. Feel free to change the color, width, and type of lines using additional arguments. As always, check the help menu for details.

Scatterplots

Open the help menu for plot. Then open the help menu for par. Here you can see all the different arguments you can specify in plots.
Create a really boring scatterplot of exercise and happiness by only specifying the data without additional arguments
Now create a prettier scatterplot by making the following changes
- Change the point type (with pch) to 16
- Change the y axis limits to cover the entire possible happiness range (0 to 100).
- Change the colors of the points (with col) to transparent(‘steelblue4’, .8), or any other color you’d like.
- Add labels for axis and main plot title
You can easily add a regression line to a plot using a combination of abline(), and lm(). Add a regression line to your plot using the following code:

# Define the model
mod <- lm(happiness ~ exercise, 
          data = hapsur)

# Add model line to the plot
abline(reg = mod,
       lty = 2)

Repeat the previous plot, but now create separate points for people in a relationship and those not in a relationship.
- Create two new dataframes called hapsur.yrel and hapsur.nrel, each containing data from only people in a relationship or not in a relationship respectively.
- Using plot(), plot the data for people in a relationship only. Make it look interesting by changing col, pch, etc.
- Using points(), add the data for people NOT in a relationship. Make the points different from the previous points!
- Add a legend indicating which points are which.

Barplots and Pirateplots

Open the help menu for barplot
Using aggregate(), calculate the mean number of drinks separately for men and women and assign the result to an object called drinks.agg
Using barplot(), create a barplot of the average drink data for men and women (using the drinks.agg object you just created)
Open the help menu for pirateplot

?pirateplot

Now create a pirateplot showing the relationship between sex and drinks
Change the look of the plot by changing the pal, point.pch, pal, point.o, and bean.o arguments
Create a pireateplot showing the relationship between sex AND relationship status on happiness. That is, include both independent variables sex and relationship in your plot. Try using the ‘basel’ or ‘google’ palette!
Create a pirateplot showing the relationship between drinks and happiness.

Creating a “balloooon plot”

You can have a lot of fun with plots. If you draw a picture by hand, you can create it in R. Let’s make a new type of plot called a balloooon plot (yes, it’s spelled baloooon, not baloon – because of the long “o”). Try running the following code a few times and see what happens. Run each line one after another to see how the plot is generated. Then try changing some of the parameters. You can change the number of balloons with N.balloons, the colors with my.palette, and the maximum size of the balloons in the line where the size object is defined

N.balloons <- 100  # Try 10, 100, 1000
my.palette <- "basel" # try "basel", "google", "nemo"

# x - locations of balloons
x.loc <- rnorm(N.balloons, mean = 100, sd = 15)

# y.loc - vertical locations of balloons
y.loc <- rnorm(N.balloons, mean = 100, sd = 10)

# size - size of the balloons
size <- runif(N.balloons, min = 0, max = 3)

# Set up the plotting space

# Remove margins
par(mar = c(0, 0, 0, 0))

plot(1, 
     xaxt = "n", 
     yaxt = "n", 
     bty = "n", 
     xlab = "", 
     ylab = "",
     xlim = c(70, 130),
     ylim = c(70, 130),
     type = "n"
)

# Add Strings
segments(x0 = x.loc + rnorm(N.balloons, mean = 0, sd = .3), 
         y0 = y.loc - (size * 1.5), 
         x1 = x.loc, 
         y1 = y.loc, 
         col = transparent("black", 1 - size / 10),
         lwd = size / 3
         )

# Add balloons
points(x.loc, 
       y.loc, 
       cex = size, # Size of the balloons
       pch = 21, 
       col = "white", # white border
       bg = piratepal(palette = my.palette, trans = .5))

Let’s make the baloooon plot look a bit more scientific by ncluding axes, gridlines, a regression line, etc. Oh, and I’ll also add a balloon for each person in the class. Run the following code and see if you can find your balloon. If you can’t find it, just run the code again!

class.initials <- c("AA", "RA", "CB", "GB", "TB", "GaBr", "EE", "VF", "SF", "SH", 
"WH", "SeHu", "LK", "TK", "LL", "SM", "JM", "LN", "KO", "NP", 
"SR", "AS", "SS", "CS", "GS", "LS", "ShSt", "KS", "SaSt", "Bv", 
"LW")

N.balloons <- length(class.initials)  
my.palette <- "basel" # try "basel", "google", "nemo"

# x - locations of balloons
x.loc <- rnorm(N.balloons, mean = 100, sd = 15)

# y.loc - vertical locations of balloons
y.loc <- x.loc + rnorm(N.balloons, mean = 0, sd = 15)

# size - size of the balloons
size <- rexp(N.balloons, rate = .7)

# Set up the plotting space

# Regular margins
par(mar = c(5, 4, 4, 1))

plot(1, 
     xlab = "Drinks", 
     ylab = "Happiness",
     main = "R Pirate drinks per day and Happiness",
     xlim = c(60, 140),
     ylim = c(60, 140),
     type = "n")

mtext("Point size indicates love of R", line = .5)

# Add gray background and gridlines
rect(-1000, -1000, 1000, 1000, col = gray(.96))

abline(h = seq(60, 140, 5), lwd = c(.5, 1), col = "white",
       v = seq(60, 140, 5))

# Add regression line

abline(lm(y.loc ~ x.loc, 
          data = as.data.frame(cbind(y.loc, x.loc))),
       lty = 2)

# Add initials below each balloon string

text(x = x.loc,
     y = y.loc - (size * 2),
     labels = class.initials,
     cex = size / 3,
     pos = 1
     )

# Add Strings
segments(x0 = x.loc + rnorm(N.balloons, mean = 0, sd = .3), 
         y0 = y.loc - (size * 1.5), 
         x1 = x.loc, 
         y1 = y.loc, 
         col = transparent("black", 1 - size / max(size)),
         lwd = size / 5
         )

# Add balloons
points(x.loc, 
       y.loc, 
       cex = size, # Size of the balloons
       pch = 21, 
       col = "white", # white border
       bg = piratepal(palette = my.palette, trans = .5))

WPA #5: Chapter 10 – Plotting