Titanic
The content above in lines 1 through 17 is called the YAML, and it is useful to set the document title and specify the output when you render the file. Rendering options include html, Word, and .pdf. Here, we will render the document into an html file with folded code that wraps. Note, too that we include a table of contents (toc) and we can apply the global options to suppress warnings and messages as well.
This text in the white sections are areas where you can insert narratives and other useful information. The backslash at the end is a carriage return – which means start a new line, otherwise this text will run together.
The gray section below is called a code chunk. This is where you will place R script.
All code chunks start and end with three tick marks — use the key just below the ESC key on a Windows keyboard.
The first line needs the {r} to indicate that this code chunk is using R script. (Quarto will run with other code types such as Python and you can include different code chunks in the same document. However, for this course, we will only use R throughout.)
You can run all of the commands just in this code chunk by clicking on the green play arrow in the upper right corner of the chunk. Go ahead and run the code below. Note, you may need to install the titanic package first using the Packages tab in the lower right quadrant of RStudio.
For this exercise, we will use the dataset that is stored in the titanic package. Note that the data are already split into a training and test datasets. We want to use all records, so the first thing we will do is the concatenate the datasets (bind_rows) into one.
# Note, I used ChatGPT for some parts of the code and to cite the parts where it was used I placed a comment on the same line as the code saying "Used ChatGPT".
# combine test and train datasets
library(titanic)
library(dplyr)
titanic <- bind_rows(titanic_train, titanic_test) # concatenation of tables Let’s understand the structure of the dataset as well as view a few of the first and last observations.
# view data and metadata
str(titanic) # look at the structure of the dataset'data.frame': 1309 obs. of 12 variables:
$ PassengerId: int 1 2 3 4 5 6 7 8 9 10 ...
$ Survived : int 0 1 1 1 0 0 0 0 1 1 ...
$ Pclass : int 3 1 3 1 3 3 1 3 3 2 ...
$ Name : chr "Braund, Mr. Owen Harris" "Cumings, Mrs. John Bradley (Florence Briggs Thayer)" "Heikkinen, Miss. Laina" "Futrelle, Mrs. Jacques Heath (Lily May Peel)" ...
$ Sex : chr "male" "female" "female" "female" ...
$ Age : num 22 38 26 35 35 NA 54 2 27 14 ...
$ SibSp : int 1 1 0 1 0 0 0 3 0 1 ...
$ Parch : int 0 0 0 0 0 0 0 1 2 0 ...
$ Ticket : chr "A/5 21171" "PC 17599" "STON/O2. 3101282" "113803" ...
$ Fare : num 7.25 71.28 7.92 53.1 8.05 ...
$ Cabin : chr "" "C85" "" "C123" ...
$ Embarked : chr "S" "C" "S" "S" ...head(titanic,3) # look at the first 3 rows of the dataset PassengerId Survived Pclass
1 1 0 3
2 2 1 1
3 3 1 3
Name Sex Age SibSp Parch
1 Braund, Mr. Owen Harris male 22 1 0
2 Cumings, Mrs. John Bradley (Florence Briggs Thayer) female 38 1 0
3 Heikkinen, Miss. Laina female 26 0 0
Ticket Fare Cabin Embarked
1 A/5 21171 7.2500 S
2 PC 17599 71.2833 C85 C
3 STON/O2. 3101282 7.9250 Stail(titanic,3) # look at the last 3 rows of the dataset PassengerId Survived Pclass Name Sex Age SibSp
1307 1307 NA 3 Saether, Mr. Simon Sivertsen male 38.5 0
1308 1308 NA 3 Ware, Mr. Frederick male NA 0
1309 1309 NA 3 Peter, Master. Michael J male NA 1
Parch Ticket Fare Cabin Embarked
1307 0 SOTON/O.Q. 3101262 7.2500 S
1308 0 359309 8.0500 S
1309 1 2668 22.3583 CNote that important variables such as \(Survived\) and \(Age\) have missing values.
summary(titanic) # look at the descriptive statistics for the numeric variables PassengerId Survived Pclass Name
Min. : 1 Min. :0.0000 Min. :1.000 Length:1309
1st Qu.: 328 1st Qu.:0.0000 1st Qu.:2.000 Class :character
Median : 655 Median :0.0000 Median :3.000 Mode :character
Mean : 655 Mean :0.3838 Mean :2.295
3rd Qu.: 982 3rd Qu.:1.0000 3rd Qu.:3.000
Max. :1309 Max. :1.0000 Max. :3.000
NA's :418
Sex Age SibSp Parch
Length:1309 Min. : 0.17 Min. :0.0000 Min. :0.000
Class :character 1st Qu.:21.00 1st Qu.:0.0000 1st Qu.:0.000
Mode :character Median :28.00 Median :0.0000 Median :0.000
Mean :29.88 Mean :0.4989 Mean :0.385
3rd Qu.:39.00 3rd Qu.:1.0000 3rd Qu.:0.000
Max. :80.00 Max. :8.0000 Max. :9.000
NA's :263
Ticket Fare Cabin Embarked
Length:1309 Min. : 0.000 Length:1309 Length:1309
Class :character 1st Qu.: 7.896 Class :character Class :character
Mode :character Median : 14.454 Mode :character Mode :character
Mean : 33.295
3rd Qu.: 31.275
Max. :512.329
NA's :1 John Jacob Astor, one of the wealthiest people in the world, was aboard the Titanic. Let’s see if he survived.
astor <- titanic[grepl("Astor", titanic$Name), ]
print(astor %>% select(Name, Survived)) Name Survived
701 Astor, Mrs. John Jacob (Madeleine Talmadge Force) 1
1094 Astor, Col. John Jacob NACol. Astor was not known to have survived nor was he confirmed as a non-survivor. The NA means he was not recovered. His wife, however, did survive.
#create a new variable that is the average age. the na.rm=TRUE tells R to skip missing values.
xage <- round(mean(titanic$Age, na.rm = TRUE), digits = 2)Rather than hardcode the value for the age, we will use inline code to print the value in the narrative. This is useful if the variable you create is dynamic, such as when you frequently update a dataset with new data. To do this, we use inline code, like this:
- The average age of all passengers on board the Titanic was 29.88.
- You will see the answer when you render the code.
Let’s run just a few more bits of analysis. Specifically, let’s see if we can answer the following questions:
1. How many passengers are male and how many female?
2. How many passengers survived and how many died?
3. How many females survived? How many died?
4. What percent of females survived?
5. What percent of survivors were male?
6. What percent of passengers were females who perished?
7. What is the distribution of passengers by fare class and embark location?
The following table shows the number of female and male passengers aboard the Titanic.
table(titanic$Sex)
female male
466 843 However, it may be nicer to show the answer as inline text vs code output:
females <- nrow(titanic[titanic$Sex == "female", ])
males <- nrow(titanic[titanic$Sex == "male", ])The following table shows the number of known survivors and non-survivors from the Titanic.
Note that many of the passengers were unaccounted (Unsure).
# convert the variable to a factor
titanic$Survived.f <- as.factor(ifelse(is.na(titanic$Survived), "Unsure", # if the value is NA, then unsure
ifelse(titanic$Survived == 0, "Did Not Survive",
"Survived")))
table(titanic$Survived.f)
Did Not Survive Survived Unsure
549 342 418 table1 <- as.data.frame(table(titanic$Survived.f)) The following table shows the number of survivors and non-survivors by gender.
table(titanic$Sex,titanic$Survived.f)
Did Not Survive Survived Unsure
female 81 233 152
male 468 109 266# use this to create the values for inline code
female_survived <- nrow(titanic[titanic$Sex == "female" & titanic$Survived.f == "Survived", ])
female_nonsurvived <- nrow(titanic[titanic$Sex == "female" & titanic$Survived.f == "Did Not Survive" , ])
# use this to create the values for inline code
male_survived <- nrow(titanic[titanic$Sex == "male" & titanic$Survived.f == "Survived", ])
male_nonsurvived <- nrow(titanic[titanic$Sex == "male" & titanic$Survived.f == "Did Not Survive" , ])Records show that among the females on board the Titanic, 233 survived and 81 did not survive.
While the prior table showed the number, this table depicts the percentage of survivors by gender.
# create contingency table
library(summarytools)
ctable(titanic$Sex,titanic$Survived.f, prop="r")Cross-Tabulation, Row Proportions
Sex * Survived.f
Data Frame: titanic
-------- ------------ ----------------- ------------- ------------- ---------------
Survived.f Did Not Survive Survived Unsure Total
Sex
female 81 (17.4%) 233 (50.0%) 152 (32.6%) 466 (100.0%)
male 468 (55.5%) 109 (12.9%) 266 (31.6%) 843 (100.0%)
Total 549 (41.9%) 342 (26.1%) 418 (31.9%) 1309 (100.0%)
-------- ------------ ----------------- ------------- ------------- ---------------# rather do this by hand so you can use inline code
total_females <- sum(titanic$Sex == "female") # count females
percentage_female_survived <- round((female_survived / total_females) * 100, digits=1)
percentage_female_nonsurvived <- round((female_nonsurvived / total_females) * 100, digits=1)Among Females on board the Titanic, 50% survived and 17.4% did not survive.
The following table shows the distribution of survivors by gender.
# insert code here. note the way this is worded.
# Column proportions
# create contingency table
library(summarytools)
ctable(titanic$Sex,titanic$Survived.f, prop="c")Cross-Tabulation, Column Proportions
Sex * Survived.f
Data Frame: titanic
-------- ------------ ----------------- -------------- -------------- ---------------
Survived.f Did Not Survive Survived Unsure Total
Sex
female 81 ( 14.8%) 233 ( 68.1%) 152 ( 36.4%) 466 ( 35.6%)
male 468 ( 85.2%) 109 ( 31.9%) 266 ( 63.6%) 843 ( 64.4%)
Total 549 (100.0%) 342 (100.0%) 418 (100.0%) 1309 (100.0%)
-------- ------------ ----------------- -------------- -------------- ---------------# rather do this by hand so you can use inline code
total_survivors <- sum(titanic$Survived.f == "Survived") # count females
percentage_male_survived <- round((male_survived / total_survivors) * 100, digits=1)
percentage_male_nonsurvived <- round((male_nonsurvived / total_survivors) * 100, digits=1)31.9% of the survivors were male.
# insert code here to complete.
# rather do this by hand so you can use inline code
percentage_female_nonsurvived <- round((female_nonsurvived / total_females) * 100, digits=1)17.4% of female passengers who were females who perished.
Note that passengers embarked at one of three locations.
S: Southampton, England
C: Cherbourg, France
Q: Queenstown, Ireland
# fix missing labels and convert to a factor
titanic$Embarked <- ifelse(titanic$Embarked=="","Unknown",titanic$Embarked)
titanic$Embarked.f <- as.factor(titanic$Embarked)
# insert code here to complete.
#head(titanic)
#table(titanic$Embarked.f, titanic$Pclass)
library(summarytools)
# When doing contingency tables ensure the categorical variables are factors
# Factors are a type of variable used to store variables with specific level of categorization
# You can also use character variables
contingencyEmbarked <- ctable(titanic$Embarked.f, as.factor(titanic$Pclass), prop="r")
contingencyEmbarkedCross-Tabulation, Row Proportions
Embarked.f * as.factor(titanic$Pclass)
------------ --------------------------- -------------- ------------- ------------- ---------------
as.factor(titanic$Pclass) 1 2 3 Total
Embarked.f
C 141 ( 52.2%) 28 (10.4%) 101 (37.4%) 270 (100.0%)
Q 3 ( 2.4%) 7 ( 5.7%) 113 (91.9%) 123 (100.0%)
S 177 ( 19.4%) 242 (26.5%) 495 (54.2%) 914 (100.0%)
Unknown 2 (100.0%) 0 ( 0.0%) 0 ( 0.0%) 2 (100.0%)
Total 323 ( 24.7%) 277 (21.2%) 709 (54.2%) 1309 (100.0%)
------------ --------------------------- -------------- ------------- ------------- ---------------# contingencyTable <- contingencyEmbarked$proportions
# print(contingencyTable[1, 1])Let’s create three plots using ggplot2. We will learn more about the capabilities of ggplot2 over this course. We will practice a few here.
The first two charts are similar – one shows the levels, and the other shows the values in percent.
library(ggplot2)
ggplot(titanic, aes(x = Sex, fill = Survived.f)) +
geom_bar(position = "dodge") +
geom_text(aes(label = after_stat(count)), stat = "count", position = position_dodge(width = 0.9), vjust = -0.25) +
labs(x = "Gender", y = "Count", fill = "Survival Status",
title = "Number of Titanic Passengers by Survival and Gender",
subtitle = "Most passengers aboard the Titanic were men who did not survive") 
This is a similar chart as above, but showing the data as a percent rather than raw values.
Note that we first need to calculate the percentage first using dplyr then we will use that data to create the chart. We also add data labels.
# summarize the data to create the percentages
pip <- titanic %>%
group_by(Sex, Survived.f) %>%
summarise(Count = n(), .groups = 'drop') %>%
mutate(Total = sum(Count), Percentage = (Count / Total) * 100) # Creates new variables within data frame.
pip# A tibble: 6 × 5
Sex Survived.f Count Total Percentage
<chr> <fct> <int> <int> <dbl>
1 female Did Not Survive 81 1309 6.19
2 female Survived 233 1309 17.8
3 female Unsure 152 1309 11.6
4 male Did Not Survive 468 1309 35.8
5 male Survived 109 1309 8.33
6 male Unsure 266 1309 20.3 # Create the dodged bar chart
ggplot(pip, aes(x = Sex, y = Percentage, fill = Survived.f)) +
geom_bar(stat = "identity", position = position_dodge(width = 0.9)) +
geom_text(aes(label = sprintf("%.0f%%", Percentage), #sprintf rounds to 0 decimals
group = Survived.f),
position = position_dodge(width = 0.9), vjust = -0.25) +
scale_y_continuous(labels = scales::percent_format()) +
labs(x = "Gender", y = "Percentage", fill = "Survival Status",
title = "Percentage of Survival by Gender on the Titanic",
subtitle = "Nearly 36% of all passengers were men who did not survive") +
theme(axis.text.y = element_blank()) # Suppress y-axis labels
Create a bar chart that shows the average age of passengers by fare class – 1st Class, 2nd Class, or 3rd Class.
# Create a file with the average ages
library(dplyr)
mean_ages <- titanic %>%
group_by(Pclass) %>%
summarise(MeanAge = round(mean(Age, na.rm=TRUE), 0))
# Create the bar chart
library(ggplot2)
ggplot(data=mean_ages, aes(x=Pclass, y=MeanAge)) +
geom_bar(stat="identity", fill="light blue", position="dodge") +
geom_text(aes(label=round(MeanAge, 1)), vjust=-0.5) +
ylim(0,45) +
labs(x = "Passenger Class", y="Average Age",
title = "Average Age by Passenger Class",
subtitle = "First class passengers were older, on average.",
caption = "Source: titanic dataset") 
Here we created the plot that shows the average age by fare class. First class Titanic passengers were, on average, older than those in second and third class.
Create and interpret two charts that address embark location and fare class.
Include both the intent, the chart, and the interpretation.
Be prepared to share your charts with the class.
# Note: Find another question and have fun choosing what else you want to make a chart of.
ggplot(titanic, aes(x = Embarked.f, fill = as.factor(Pclass))) +
geom_bar(position = "dodge") +
geom_text(aes(label = after_stat(count)), stat = "count", position = position_dodge(width = 0.9), vjust = -0.25) +
labs(x = "Embarked", y = "Count", fill = "Fare Class",
title = "Embark Location and Fare Class",
subtitle = "Half the passengers from Cherbourg, France were first class") 
About half of the passengers from Cherbourg, France, were first class. Neither of the other two embarking locations had as high of a percentage of passengers in the first class. Maybe a higher percentage of the population in Cherbourg is wealthy.
more_titanic <- titanic[!is.na(titanic$Age), ] # Used ChatGPT
# Note: Find another question and have fun choosing what else you want to make a chart of.
group_num <- 2 # Used ChatGPT (I was changing the group number often and decided to use two so I could refer to the groups as younger and older)
# I chose to use a quantile instead of intervals so the median age of all the passengers aboard the Titanic would determine what is considered to be younger or older passengers.
breaks <- quantile(more_titanic$Age, probs = seq(0, 1, length.out = group_num + 1)) # Used ChatGPT
rounded_breaks <- round(breaks, 1) # Used ChatGPT
interval_labels <- paste(head(rounded_breaks, -1), "-", tail(rounded_breaks, -1)) # Used ChatGPT
age_groups <- cut(more_titanic$Age, breaks = breaks, labels = interval_labels, include.lowest = TRUE) # Used ChatGPT
more_titanic$ageGroup <- age_groups
ggplot(more_titanic, aes(x = Embarked.f, fill = as.factor(ageGroup))) +
geom_bar(position = "dodge") +
geom_text(aes(label = after_stat(count)), stat = "count", position = position_dodge(width = 0.9), vjust = -0.25) +
labs(x = "Embarked", y = "Count", fill = "Passenger Ages",
title = "Embark Location and Age",
subtitle = "More older passengers than younger passengesrs came from Cherbourg, France")
There are slightly more older people from Cherbourg, which may be because many from this area are in the first class and the first class has a higher average age. There is about an equal amount of younger and older people from Queenstown. There are slightly more younger people from Southampton.