# Read in the training data (passengers with whether or not they survived)
train <- read.csv("titanic_train.csv",stringsAsFactors = TRUE)
# Read in the test data (passengers without the Survived column,
# which we are trying to predict)
test <- read.csv("titanic_test.csv",stringsAsFactors = TRUE)

# Load the package (R version of a Python module) rpart,
# which has functions (commands) for decision trees.
library(rpart)
# "Fit" the model (the decision tree) to the training data.
# We will use the information in columns Pclass, Sex, Age, SibSp, 
# Parch, Fare, Embarked (independent variables) to make our decision tree.
# We are trying to predict Survived (dependent variable).
fit <- rpart(Survived ~ Pclass + Sex + Age + SibSp + Parch +Fare + Embarked, data = train, method = "class")
# Plot the decision tree (the model).
plot(fit)
text(fit)

# To plot a fancier picture of the decision tree:
# If you haven't already, install the package rpart.plot
install.packages('rpart.plot')
# Load the package into our working space
library(rpart.plot)
# Make a fancier plot of the decision tree.
prp(fit)

# Use the decision tree to make predictions of who in the test data survived
Prediction <- predict(fit,test,type = 'class')

# Create a new dataframe with only the PassengerId and our prediction.
submit <-data.frame(PassengerId = test$PassengerId, Survived = Prediction)
# Write this new dataframe to a csv file, for submission to Kaggle.
write.csv(submit, file = "decisiontree.csv", row.names = FALSE)