Topics Final Project: Data Science with R-Studio

The first think we need to do when working with a data set is make sure the data is in the same place as our code.

click “session”

“set working directory” > “change directory”

Then choose the folder in which you saved the data set.

You are now ready to import the data and take a look at the variables contained in the set.

 sleep=read.csv("SleepStudy.csv", header=TRUE)
head(sleep)

Next, let’s make sure we have the tools we need to analyze the data.

We will start with a tool to help us better work with the data set.

#only need to run this once
install.packages("dplyr", dependencies=TRUE)

If you have already installed dplyr before, please just use the following.

library(dplyr)

Attaching package: <U+393C><U+3E31>dplyr<U+393C><U+3E32>

The following objects are masked from <U+393C><U+3E31>package:stats<U+393C><U+3E32>:

    filter, lag

The following objects are masked from <U+393C><U+3E31>package:base<U+393C><U+3E32>:

    intersect, setdiff, setequal, union

We will use ggplot2 for graphics. Run the following command to load the ggplot package (we have already installed it).

#only need to run this once
install.packages("ggplot2")

If you have already installed ggplot2, please use the following.

library(ggplot2)

Exploratory Analysis: We will take a look at the data using graphical and numerical tools.

Categorical Data Analysis

LarkOwl vs. Alcohol use, for example, will give us a sense of whether alcohol use is related to when students are awake/asleep. We can begin this exploration with a bar graph.

ggplot(sleep, aes(AlcoholUse, fill=LarkOwl))+geom_bar(position = "dodge")

We can use the following command to group

sleep %>%
  group_by(AlcoholUse,LarkOwl) %>%
  summarize(n())

**Optional: Put the results into a two way table:

table=matrix(c(6,20,8,2,10,4,19,59,5,14,74,32),nrow=4,ncol=3, byrow=TRUE)
rownames(table)=c("Abstain","Heavy","Light","Moderate")
colnames(table)=c("Lark","Neither","Owl")
table

         Lark Neither Owl
Abstain     6      20   8
Heavy       2      10   4
Light      19      59   5
Moderate   14      74  32

Quantitative Data Analysis

Histogram of Sleep Quality

qplot(sleep$PoorSleepQuality,
      geom="histogram", 
      binwidth = 1,
      fill=I("tomato3"))

Boxplot of Sleep Quality

ggplot(sleep, aes(x="",y = PoorSleepQuality)) +
        geom_boxplot()+
  ggtitle("Sleep Quality Score")

Summary of Sleep Quality

summary(sleep$PoorSleepQuality)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   4.000   6.000   6.257   8.000  18.000 

Standard Deviation of Sleep Quality

sd(sleep$PoorSleepQuality)

[1] 2.919761

Histogram of Anxiety

qplot(sleep$AnxietyScore,
      geom="histogram", 
      binwidth = 1,
      fill=I("khaki1"))

Boxplot of Anxiety

ggplot(sleep, aes(x="",y = AnxietyScore)) +
        geom_boxplot()+
  ggtitle("Anxiety Score College Students")

Summary of Anxiety

summary(sleep$AnxietyScore)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  0.000   1.000   4.000   5.372   8.000  26.000 

Standard Deviation of Anxiety

sd(sleep$AnxietyScore)

[1] 5.197327

**Optional: Multivariable Plots

Quality of Sleep by type Lark, Neither, Owl

ggplot(sleep, aes(x=LarkOwl,y = PoorSleepQuality)) +
        geom_boxplot()+
  ggtitle("Poor Sleep Quality of College Students")

**Optional: Anxiety based on Alcohol use

ggplot(sleep, aes(x=AlcoholUse,y = AnxietyScore)) +
        geom_boxplot()+
  ggtitle("Anxiety Score College Students")

Two Variable Analysis

Here we will look at the relationship between two quantitative variables

ggplot(sleep, aes(x=PoorSleepQuality, y=AnxietyScore)) +
  geom_point(size=2)+
  ggtitle("Anxiety vs. Poor Sleep Score")

Calculate the Correlation

cor(sleep$PoorSleepQuality,sleep$AnxietyScore)

[1] 0.4513103

Fit a line to the data and add it to the plot

ggplot(sleep, aes(x=PoorSleepQuality, y=AnxietyScore)) +
  geom_point(size=2)+
  stat_smooth(method="lm",col="red",se=FALSE)+
  ggtitle("Anxiety vs. Poor Sleep Score")

Find the equation of the best fit line

#Fit the linear model
sleep.lm=lm(sleep$AnxietyScore~sleep$PoorSleepQuality)
coefficients(sleep.lm) 

           (Intercept) sleep$PoorSleepQuality 
             0.3450103              0.8033559 

The slope is 0.803 and the intercept is 0.345

So the model predicting Anxiety score from sleep quality is:

Anxiety = 0.803*PoorSleep + 0.345

---

**Optional: Color code by a category

ggplot(sleep, aes(x=PoorSleepQuality, y=AnxietyScore,   color=AlcoholUse)) +
  geom_point(size=2)+
  ggtitle("Anxiety vs. Poor Sleep by Alcohol Use")