R_Week_3
Ian Costello
8/2/2020
R Bridge Week 3 - Titanic Surviors
The R.M.S. Titanic has captivated the world since it sank the night of April 14, 1912. Scrolling through the data sets, even seeing the word “Titanic” was so evocative I had to select it even just to see what it contained.
Analytical Question
The Titanic disaster was a turning point for marine transportation even today. Many rules, laws, and safety procedures that stemmed from investigation and recommendations following the sinking are still in place today. The question I wish to address using the current data set is:
What did the typical survivor and victim look like?
Perhaps obviously, my assumption before doing much analysis is that women, children, and 1st class passengers would be the most likely survivor and older, male, and 3rd class passengers the most likely victim, following in the “women and children first” trope. I wish this data set had additional categories, such as crew or passenger status, national origin, and room and deck assignments.
Summary Statistics
Downloading the data set and re-uploading it to my GitHub successfully pulling from my account.
TiSurURL <- "https://raw.githubusercontent.com/iscostello/Rbridgedata/master/TitanicSurvival.csv"
TiSur <- read.csv(file = TiSurURL, header = TRUE, sep = ",")
head(TiSur)## X survived sex age passengerClass
## 1 Allen, Miss. Elisabeth Walton yes female 29.0000 1st
## 2 Allison, Master. Hudson Trevor yes male 0.9167 1st
## 3 Allison, Miss. Helen Loraine no female 2.0000 1st
## 4 Allison, Mr. Hudson Joshua Crei no male 30.0000 1st
## 5 Allison, Mrs. Hudson J C (Bessi no female 25.0000 1st
## 6 Anderson, Mr. Harry yes male 48.0000 1stsummary(TiSur)## X survived sex age
## Length:1309 Length:1309 Length:1309 Min. : 0.1667
## Class :character Class :character Class :character 1st Qu.:21.0000
## Mode :character Mode :character Mode :character Median :28.0000
## Mean :29.8811
## 3rd Qu.:39.0000
## Max. :80.0000
## NA's :263
## passengerClass
## Length:1309
## Class :character
## Mode :character
##
##
##
## The summary statistics for the entire data set do not offer much insight into my question. However, there are some interesting observations.
- The data set is incomplete
- Titanic had 2,208 people on board, of those 1,317 were passengers
- This data set has only 1,309 records, so assuming this is looking at only passengers
- The people in this data set trends very young
- Median age of 29
- NULL age records (263) are abundant enough to perhaps affect these statistics
- Both data elements for survival and sex are stored as characters, so the summary function will not sort them automatically
Data Wrangling
TiSur$survivedNew <- ifelse(TiSur$survived == "yes", TRUE, FALSE)
TiSur$sexNew <- ifelse(TiSur$sex == "female", TRUE, FALSE)
TiSur$passClassNew <- as.factor(ifelse(TiSur$passengerClass == "1st", 1,
ifelse(TiSur$passengerClass == "2nd", 2,
ifelse(TiSur$passengerClass == "3rd", 3,
NA ))))
summary(TiSur)## X survived sex age
## Length:1309 Length:1309 Length:1309 Min. : 0.1667
## Class :character Class :character Class :character 1st Qu.:21.0000
## Mode :character Mode :character Mode :character Median :28.0000
## Mean :29.8811
## 3rd Qu.:39.0000
## Max. :80.0000
## NA's :263
## passengerClass survivedNew sexNew passClassNew
## Length:1309 Mode :logical Mode :logical 1:323
## Class :character FALSE:809 FALSE:843 2:277
## Mode :character TRUE :500 TRUE :466 3:709
##
##
##
## The two columns “survived” and “sex” are stored in the original data set as characters. In order to summarize the data with the summary function, and, indeed, be more useful for other analytics I created two new columns. At first, I started to rename the values as numbers “1” for a “yes” response, “0” for “no.” Because these values are binary I could store them as a boolean, that is a “True/False”.
Perhaps the easiest solution is to simply convert sex and survivor status as a factor, as I’ve done for class. Seeing class in this way and compared to the passenger summaries from the internet, I see that the 3rd class passengers add up exactly to the official tally. This data set is missing 7 2nd class passengers and 1 1st class passenger.
TiSurOnly <- subset(TiSur, TiSur$survivedNew == TRUE)
TiSurVictim <- subset(TiSur, TiSur$survivedNew == FALSE)Scatter Plot of Passengers by Age and Class
Probably not the best visual for a scatter plot. The data set is pretty light on continuous data, so these lines are a bit tricky to interperate. I much prefer the histograms below to demonstrate these data.
library(ggplot2)
ggplot(TiSur, aes(x=age, y=passClassNew, color=survivedNew)) +
geom_point()## Warning: Removed 263 rows containing missing values (geom_point).
Box Plot of Age by Survivor Status
library(ggplot2)
ggplot(TiSur, aes(x=survivedNew, y=age)) +
geom_boxplot()## Warning: Removed 263 rows containing non-finite values (stat_boxplot).
Histogram of Survivors ONLY by Sex (Gender)
library(ggplot2)
ggplot(TiSurOnly, aes(x=age, color=sex)) +
geom_histogram(fill="white", alpha=0.1, position="stack")## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.## Warning: Removed 73 rows containing non-finite values (stat_bin).
Histogram of Victims ONLY by Passenger Class Status
library(ggplot2)
ggplot(TiSurVictim, aes(x=age, color=sex)) +
geom_histogram(fill="white", alpha=0.1, position="stack")## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.## Warning: Removed 190 rows containing non-finite values (stat_bin).
Histogram of ALL Passengers by Age and Survivor Status
library(ggplot2)
ggplot(TiSur, aes(x=age, color=survivedNew)) +
geom_histogram(fill="white", alpha=0.1, position="stack")## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.## Warning: Removed 263 rows containing non-finite values (stat_bin).
Conclusions
Though the data set was fairly limited, I was able to draw a few conclusions based on the summary statistics and graphics created.
- 3rd class passengers were more likely victims
- There were many more 3rd class passengers than 1st or 2nd
- While these data are not in this set, it is well-known that 3rd class passengers were disadvantaged by their location on the ship and, likely in part, social status
- Women were more likely survivors - No suprise that my initial guess would pan out.
- Age does not appear determinative - While the box plot shows that survivors were slightly younger, it has had a number of outliers, which are the oldest passengers in the set.
The typical survivor was most likely 1st class and a woman, while the typical victim was likely 3rd class an a man.