library(dplyr)
######################################
# 1. Load Data
######################################
# Set Working Directory
setwd("/Users/oindrilasen/WORK_AREA/Data Science/kaggle/Titanic")
# Read train.csv data file
titanic_clean<-read.csv("train.csv",
                        header = TRUE,
                        na.strings = "",
                        stringsAsFactors = FALSE)

glimpse(titanic_clean)
######################################
# 2. Data Wrangling and Cleaning
######################################
# Check which variables are Factors
sapply(titanic_clean, function(x) length(unique(x)))
# Transforming categorical Variables to factors:
to_factor <- c(
  'Survived',
  'Pclass',
  'Sex',
  'Embarked'
)
for (col in to_factor) {
  titanic_clean[[col]] <- factor(titanic_clean[[col]])
}
# check for NA values
sapply(titanic_clean, function(x) sum(is.na(x)))

# Convert Age column to Numeric
titanic_clean$Age <- as.integer((titanic_clean$Age))
# Relace NA values for Age with the Median
titanic_clean$Age[is.na(titanic_clean$Age)] <- mean(titanic_clean$Age,na.rm = TRUE)
# Replace Cabin# with None for NA records
titanic_clean$Cabin[is.na(titanic_clean$Cabin)] <- "None"
# Check for Embarked variable
table(titanic_clean$Embarked)
# Relace the Embarked value with the most common value i.e S
titanic_clean$Embarked[is.na(titanic_clean$Embarked)] <- "S"
# Convert Fare column to Integer
titanic_clean$Fare <- as.integer((titanic_clean$Fare))
# Relace NA values for Age with the Median
titanic_clean$Fare[is.na(titanic_clean$Fare)] <- median(titanic_clean$Fare,na.rm = TRUE)
# Again check for NA values
sapply(titanic_clean, function(x) sum(is.na(x)))

# Change the levels to meaningful values
# 1. Pclass
levels(titanic_clean$Pclass)[levels(titanic_clean$Pclass)== "1"] <- "1st Class"
levels(titanic_clean$Pclass)[levels(titanic_clean$Pclass)== "2"] <- "2nd Class"
levels(titanic_clean$Pclass)[levels(titanic_clean$Pclass)== "3"] <- "3rd Class"

# 2. Embarked
levels(titanic_clean$Embarked)[levels(titanic_clean$Embarked)== "C"] <- "Cherbourg"
levels(titanic_clean$Embarked)[levels(titanic_clean$Embarked)== "Q"] <- "Queenstown"
levels(titanic_clean$Embarked)[levels(titanic_clean$Embarked)== "S"] <- "Southampton"
######################################
# 3. Adding New Features
######################################
# Add new feature Fare_Group
summary(titanic_clean$Fare)
titanic_clean$Fare_Group <-factor(ifelse(titanic_clean$Fare >= 0 & titanic_clean$Fare <= 15, "Low",
                                         ifelse(titanic_clean$Fare > 15 & titanic_clean$Fare <=100, "Medium",
                                                ifelse(titanic_clean$Fare >100 ,"High",NA
                                                ))))

# Add new feature Age_Group
summary(titanic_clean$Age)
titanic_clean$Age_Group <-factor(ifelse(titanic_clean$Age<= 3, "Baby",
                                 ifelse(titanic_clean$Age> 3 & titanic_clean$Age<=12, "Kid",
                                 ifelse(titanic_clean$Age> 12 & titanic_clean$Age<=18, "Teen",
                                 ifelse(titanic_clean$Age> 18, "Adult",NA
                                  ))))
)
# Add new feature with_family
titanic_clean$with_family <-factor(ifelse(titanic_clean$Parch == 0 & titanic_clean$SibSp ==0, "no","yes"))
View(titanic_clean)