Titanic App

Project for Coursera Developing Data Products

Lik Sin
Coursera Participant

Titanic App

Choose your economic status (class), sex and age.
This application will tell you if you have survived the Titanic.

Instructions

  1. Go to https://liksin.shinyapps.io/Project
  2. Wait for the app to load. App is ready when input parameters at the main panel shows user selection.
  3. Select your economic status (class) on the side panel
  4. Select your gender on the side panel
  5. Select your age on the side panel
  6. On the 'Input parameters' tab of the main panel, check that your selections are correct
  7. Based on your selected parameters, the app run them against a random forest prediction model and generate your survivability
  8. Check out the 'Server calculations and outputs' tab for details on the cross validation of the prediction model
  9. Check out 'Documentation and Codes' tab for instructions and full documentation of the code used

How it works

  • Data from Titanic dataset is first pre-processed
data=data.frame(Titanic)
flatdata=NULL
for(i in 1:4)
{
    flatdata=cbind(flatdata,rep(as.character(data[,i]),data$Freq))
}
flatdata=data.frame(flatdata)
names(flatdata)=names(data[1:4])
  • And split into training and test data sets
set.seed(150)
inTrain = createDataPartition(flatdata$Survived, p = 0.70,list=FALSE)
trainingset = flatdata[inTrain,]
testset = flatdata[-inTrain,]

Modelling and results

  • Generating random forest prediction model
modFit=train(Survived ~ ., data=trainingset, model="rf")
  • Applying prediction model to test data set
predval <- predict(modFit, newdata=testset)
table(predval,testset$Survived)

##        
## predval  No Yes
##     No  439 130
##     Yes   8  83