# Load the training data into a dataframe, with all strings being read in as factors (categories for data).
train <- read.csv("titanic_train.csv",stringsAsFactors = TRUE)
# The top line of the output of str() gives the number of rows of data in the dataframe.
str(train)

# Make a subset of the data that is only the female passengers on the ship.
women <- subset(train,subset = Sex == "female")
# Display this dataframe to see how many rows, and hence women were on the ship.
str(women)

# An alternative way to get the count for each factor in the column. 
summary(train$Sex)

# Convert the counts into proportions (probabilities)
prop.table(table(train$Sex))

# Display the proportion of passengers having each class of ticket.
prop.table(table(train$Pclass))

# Display the proportion of passengers that have a specific sex and ticket class.
prop.table(table(train$Sex,train$Pclass))

# Display the proportion of passengers of each sex and survival outcome (1 = survived, 0 = died)
prop.table(table(train$Sex,train$Survived))

# Display the proportion of passengers with each ticket class and survival outcome
prop.table(table(train$Pclass,train$Survived))

# We will now try to predict whether the passengers in the test dataset survived or not.
# Load the test data into a dataframe.
test <- read.csv("titanic_test.csv",stringsAsFactors = TRUE)
# Create a new column in this dataframe for our predictions.  To start, we fill it with all 0's.
test$Survived <- rep(0, 418)
# We will predict all the women survived and all the men died.  
# Set the Survived column to 1 if the Sex column of that row is female.
test$Survived[test$Sex == 'female'] <- 1
# To submit the predictions to Kaggle, we need a csv file with exactly two columns:  PassengerId and Survived.
# We first make a dataframe containing just these columns.
submit <- data.frame(test$PassengerId,test$Survived)
# and write the dataframe to a csv file.
write.csv(submit, file = "submission.csv", row.names = FALSE)