data <- read.csv("Data.csv")
str(data)

data <- data %>% mutate_if(is.character, as.factor)
data$SeniorCitizen <- as.factor(data$SeniorCitizen)
str(data)

data

summary(data)

check.duplicate <- data.frame(
                   row_of_data        = data %>% nrow (),
                   row_of_unique.data = data %>% distinct() %>% nrow())
check.duplicate

ColNA <- colnames(data)[ apply(data, 2, anyNA) ] # Melihat Kolom dengan daya yang hilang
ColNA

data <- na.omit(data)
data

bp1 <- data %>% 
  select(MonthlyCharges) %>% 
  ggplot(aes(MonthlyCharges)) + geom_boxplot()

bp2 <- data %>% 
  select(TotalCharges) %>% 
  ggplot(aes(TotalCharges)) + geom_boxplot()

bp3 <- data %>% 
  select(tenure) %>% 
  ggplot(aes(tenure)) + geom_boxplot()


subplot4 <- ggarrange(bp1, bp2, bp3, 
                      ncol = 1, nrow = 3,
                      common.legend = TRUE, legend = "bottom")
subplot4

# Membuat ggplot
ggpGender       <- ggplot(data, aes(x = gender, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpSenior       <- ggplot(data, aes(x = SeniorCitizen, fill = Churn)) + theme_minimal() + 
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) + 
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpPartner      <- ggplot(data, aes(x = Partner, fill = Churn)) + theme_minimal() + 
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpDependents   <- ggplot(data, aes(x = Dependents, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpPhoneService <- ggplot(data, aes(x = PhoneService, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpLines        <- ggplot(data, aes(x = MultipleLines, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpInternet     <- ggplot(data, aes(x = InternetService, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpSecurity     <- ggplot(data, aes(x = OnlineSecurity, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpBackup       <- ggplot(data, aes(x = OnlineBackup, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpProtection   <- ggplot(data, aes(x = DeviceProtection, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpSupport      <- ggplot(data, aes(x = TechSupport, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpTV           <- ggplot(data, aes(x = StreamingTV, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpMovies       <- ggplot(data, aes(x = StreamingMovies, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpContract     <- ggplot(data, aes(x = Contract, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpBilling      <- ggplot(data, aes(x = PaperlessBilling, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpMethod       <- ggplot(data, aes(x = PaymentMethod, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single")) +
                   theme(axis.text.x = element_text(angle=-45, vjust=1, hjust=0))

ggpChurn        <- ggplot(data, aes(x = Churn, fill = Churn)) + theme_minimal() +
                   geom_bar(alpha = 0.8, position = position_dodge(preserve = "single"))

# Menggabungkan ggplot
figure1 <- ggarrange(ggpGender, ggpSenior, ggpPartner, ggpDependents,
                    labels = c("Gender", "Senior", "Partner", "Dependents"),
                    ncol = 2, nrow = 2)

figure2 <- ggarrange(ggpPhoneService, ggpLines, ggpInternet, ggpSecurity,
                    labels = c("PhoneService", "Lines", "Internet", "Security"),
                    ncol = 2, nrow = 2)

figure3 <- ggarrange(ggpBackup, ggpProtection, ggpSupport, ggpTV,
                    labels = c("Backup", "Protection", "Support", "TV"),
                    ncol = 2, nrow = 2)

figure4 <- ggarrange(ggpMovies, ggpContract, ggpBilling,ggpMethod,
                    labels = c("Movies", "Contract", "Billing", "Method"),
                    ncol = 2, nrow = 2)

ggpTenure  <- ggplot(data, aes(x=tenure, fill = Churn, color=Churn)) +
              geom_histogram(alpha=0.5)
ggpMonthly <- ggplot(data, aes(x=MonthlyCharges, fill = Churn, color=Churn)) +
              geom_histogram(alpha=0.5)
ggpTotal   <- ggplot(data, aes(x=TotalCharges, fill = Churn, color=Churn)) +
              geom_histogram(alpha=0.5)

figure5    <- ggarrange(ggpTenure, ggpMonthly, ggpTotal,
                        labels = c("Tenure", "Monthly", "Total"),
                        ncol = 1, nrow = 3)
figure5

DataNum <- dplyr::select_if(data, is.numeric)
Correlation <- cor(DataNum) %>% corrplot(method = "color")

data$gender_d <- ifelse(data$gender != "Male", 1,0 )
data$Partner_d <- ifelse(data$Partner == "Yes", 1,0 )
data$Dependents_d <- ifelse(data$Dependents == "Yes", 1,0 )
data$PaperlessBilling_d <- ifelse(data$PaperlessBilling == "Yes", 1,0 )
data$PhoneService_d <- ifelse(data$PhoneService == "Yes", 1,0 )
data$MultipleLines_d <- ifelse(data$MultipleLines == "Yes", 1,0 )
data$OnlineSecurity_d <- ifelse(data$OnlineSecurity == "Yes", 1,0 )
data$OnlineBackup_d <- ifelse(data$OnlineBackup == "Yes", 1,0 )
data$DeviceProtection_d <- ifelse(data$DeviceProtection == "Yes", 1,0 )
data$TechSupport_d <- ifelse(data$TechSupport == "Yes", 1,0 )
data$StreamingTV_d <- ifelse(data$StreamingTV == "Yes", 1,0 )
data$StreamingMovies_d <- ifelse(data$StreamingMovies == "Yes", 1,0 )
data$has_InternetService_d <- ifelse(data$InternetService != "No", 1, 0)

data_tres_levels <- data %>% 
        select(InternetService,Contract,PaymentMethod)

data_tres_levels_d <-dummy_cols(data_tres_levels)

data_tres_levels_d<- data_tres_levels_d %>% 
                        select(-c(InternetService,Contract,PaymentMethod))

attach(data)
data_final <- data.frame(customerID,gender_d,SeniorCitizen,Partner_d,Dependents_d,tenure,
                         PaperlessBilling_d,PhoneService_d,MultipleLines_d,
                         OnlineSecurity_d,OnlineBackup_d,DeviceProtection_d,
                         TechSupport_d,StreamingTV_d,StreamingMovies_d,has_InternetService_d,
                         data_tres_levels_d,MonthlyCharges,TotalCharges,Churn)

data$Churn <- ifelse(data$Churn=='Yes',1,0 )
data_surv <- Surv(data$tenure, data$Churn)

plot(data_surv, xlab="Months", ylab="Survival Function", main="Kaplan-Meier's Survival Function", col=1:3)

fit <- survfit(data_surv ~ gender_d, data = data)
ggsurvplot(fit, data = data, 
           pval = TRUE, 
           conf.int = TRUE,
           )

fit <- survfit(data_surv ~ SeniorCitizen, data = data)
ggsurvplot(fit, data = data, 
           pval = TRUE, 
           conf.int = TRUE,
           )

m1 <- survfit(Surv(tenure, Churn) ~ gender, data = data)
sm1 <- surv_summary(m1, data = data)
sm1

ggsurvplot(m1, data = data,
           pval = TRUE, conf.int = TRUE,
           risk.table = "abs_pct",
           risk.table.col = "strata",
           linetype= "strata",
           break.time.by = 12, 
           surv.median.line = "hv",
           ncensor.plot = TRUE,
           ggtheme = theme_bw(),
           palette = c("#2E9FDF", "#9B59B6"))