Analysis of titanic dataset

Age distribution

Laura Zamora, Helen Villca & Laura Ramirez

Laura Zamora, Helen Villca & Laura Ramirez

For each question, write the corresponding R command & attach the relevant graphs. 1) Select import dataset as csv from https://goo.gl/At238b 2) Load into R into a variable called T3 3) Build a new dataset, titanic, by selecting these features: survived, embarked, age, sex, sibsp, parch and fare. 4) Perform a statistical analysis of the titanic dataset. 5) Survived is the dependent variable, find its proportion in the dataset. 6) Remove NAs if any. 7) Make Survived embarked and sex as factors. 8) Find the correlation matrix between survival and the other features. 9) Plot survival with other features to see if any correlation exists 10) Set a seed to 1000 and use it to split the dataset into 80% training, 20% testing 11) Build your training (till index 1046) and test (till index 1308) datasets 12) Download the rpart for “Recursive Partitioning and Regression Trees” if it is not in your R base and bring it in your environment. 13) Train your learner on the training dataset and save your model in a variable fit fit <- rpart(Survived ~ Sex + Age + SibSp + Parch + Fare + Embarked, data=train, method="class") 14) Plot your regression tree and save plot into an image file 15) You must have obtained a plot that is not very readable. In order to get a better tree, download the following packages and bring them into your environment: rattle, rpart.plot and RColorBrewer 16) Type the R command fancyRpartPlot(fit) to obtain a more readable plot 17) Examine the tree you obtained. What is the most “important” feature over which the tree first split? Do you agree with the Titanic slogan “ Women and children first!” 18) Use the predict function with your model fit to make predictions on the test dataset and save it in a variable Prediction: Prediction <- predict(fit, test, type = "class") 19) Save the results in a data frame Results that will have two columns 1: PassengerSex =test$Sex, and 2: Survived = Prediction 20) Save your data frame in a .csv file by using write.csv(Results, file = "Titanicdtree.csv", row.names = FALSE) 21) Submit your R code as a link to Rpubs and Titanicdtree.csv file on Blackboard by Tuesday, April 13th 2024.

Problems 8.1, 8.2, 8.3, and 8.7 from Kuhn and Johnson