Packages and Data

library(tidyverse)
library(ggplot2)
library(patchwork)
library(plotly)
data <- read_csv("Fraudulent_E-Commerce_Transaction_Data.csv")

Clean Data

cleanData <- function(dataframe, colsToKeep) {
  cleanedData <- dataframe %>%
    rowid_to_column("ID") %>%
    select(colsToKeep) %>%
    rename(
     Transaction_Amount=`Transaction Amount`,
     Transaction_Date=`Transaction Date`,
     Payment_Method=`Payment Method`,
     Product_Category=`Product Category`,
     Customer_Location=`Customer Location`,
     Customer_Age=`Customer Age`,
     Device_Used=`Device Used`,
     Is_Fraudulent=`Is Fraudulent`,
     Account_Age_Days=`Account Age Days`,
     Transaction_Hour=`Transaction Hour`
   )
  return(cleanedData)
}
colsToKeep <- c("ID", "Transaction Amount", "Transaction Date", "Payment Method", "Product Category", "Quantity", "Customer Location", "Customer Age", "Device Used", "Is Fraudulent", "Account Age Days", "Transaction Hour")
cleanedData <-cleanData(data, colsToKeep = colsToKeep)
## Warning: Using an external vector in selections was deprecated in tidyselect 1.1.0.
## ℹ Please use `all_of()` or `any_of()` instead.
##   # Was:
##   data %>% select(colsToKeep)
## 
##   # Now:
##   data %>% select(all_of(colsToKeep))
## 
## See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

Transaction Amount Over Time Figure 1

cleanedData %>%
  ggplot(aes(x = Transaction_Date, y = Transaction_Amount)) +
  geom_line() +
  labs(
    title = "Transaction Amount Over Time",
    x = "Date",
    y = "Transaction Amount (Dollars)"
  )

January data Figure 2

# Convert date column in cleanedData to date type
cleanedData$Transaction_Date <- as.Date(cleanedData$Transaction_Date)
startDate <- as.Date("2024-01-01")
endDate <- as.Date("2024-01-31")
januaryData <- cleanedData %>%
  filter(Transaction_Date >= startDate, Transaction_Date <= endDate)
  # Plot January data
januaryData %>%
  ggplot(aes(x = Transaction_Date, Transaction_Amount)) +
  geom_line() +
  labs(
    title = "Transactions from January 1, 2024 to January 31, 2024",
    x = "Date",
    y = "Transaction Amount (dollars)"
  )

January 26th data Figures 3 and 4

# Filter data for January 26th
j26Data <- januaryData %>%
  filter(Transaction_Date == "2024-01-26")

# Calculate proportions of fraudulent transactions for Product_Category
product_proportions <- j26Data %>%
  group_by(Product_Category) %>%
  summarise(fraudulent_proportion = sum(Is_Fraudulent == 1) / n())

# Calculate proportions of fraudulent transactions for Payment_Method
payment_proportions <- j26Data %>%
  group_by(Payment_Method) %>%
  summarise(fraudulent_proportion = sum(Is_Fraudulent == 1) / n())

# Create plot for Product_Category
fig1 <- ggplot(product_proportions, aes(x = Product_Category, y = fraudulent_proportion, fill = Product_Category)) +
  geom_bar(stat = "identity") +
  labs(title = "Relative Frequencies of Fraudulent Transactions by Product Category",
       x = "Product Category",
       y = "Proportion of Fraudulent Transactions",
       fill = "Product Category") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

# Create plot for Payment_Method
fig2 <- ggplot(payment_proportions, aes(x = Payment_Method, y = fraudulent_proportion, fill = Payment_Method)) +
  geom_bar(stat = "identity") +
  labs(title = "Relative Frequencies of Fraudulent Transactions by Payment Method",
       x = "Payment Method",
       y = "Proportion of Fraudulent Transactions",
       fill = "Payment Method") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

fig1 ##Figure 3

fig2 ## Figure 4

Health and Beauty debit card transactions on January 26th

health_beauty_credit <- j26Data %>%
  filter(Product_Category == 'health & beauty' & Payment_Method == 'debit card')
num_transactions <- nrow(health_beauty_credit)
print(num_transactions)
## [1] 826
num_fraudulent_transactions <- sum(health_beauty_credit$Is_Fraudulent == 1)
percent <- (num_fraudulent_transactions / num_transactions) * 100
print(percent)
## [1] 4.116223

Health and Beauty debit card transactions over entire date range

health_beauty_debit <- cleanedData %>%
  filter(Product_Category == 'health & beauty' & Payment_Method == 'debit card')
num_transactions <- nrow(health_beauty_debit)
print(num_transactions)
## [1] 73475
num_fraudulent_transactions <- sum(health_beauty_debit$Is_Fraudulent == 1)
percent <- (num_fraudulent_transactions / num_transactions) * 100
print(percent)
## [1] 5.007145

Debit card fraud over entire date range

debitCardData <- cleanedData %>%
  filter(Payment_Method == 'debit card') %>%
  summarise(percent = mean(Is_Fraudulent == 1, na.rm = TRUE) * 100)
debitCardData
## # A tibble: 1 × 1
##   percent
##     <dbl>
## 1    5.02
credit <- cleanedData %>%
  filter(Payment_Method == 'credit card')
num_transactions <- nrow(credit)
print(num_transactions)
## [1] 368429
num_fraudulent_transactions <- sum(credit$Is_Fraudulent == 1)
percent <- (num_fraudulent_transactions / num_transactions) * 100
print(percent)
## [1] 4.970564

Age vs Is_Fraudulent Figure 5

ageData <- cleanedData %>%
  select(Customer_Age, Is_Fraudulent)
summary_data <- ageData %>%
  group_by(Customer_Age, Is_Fraudulent) %>%
  summarise(count = n()) %>%
  filter(!is.na(Customer_Age))
## `summarise()` has grouped output by 'Customer_Age'. You can override using the
## `.groups` argument.
# Plot as a bar plot
plot_ly(summary_data, x = ~Customer_Age, y = ~count, color = ~factor(Is_Fraudulent), type = "bar") %>%
  layout(title = "Fraud by Age", xaxis = list(title = "Customer Age"), yaxis = list(title = "Count"), barmode = "group")
## Warning in RColorBrewer::brewer.pal(N, "Set2"): minimal value for n is 3, returning requested palette with 3 different levels
## Warning in RColorBrewer::brewer.pal(N, "Set2"): minimal value for n is 3, returning requested palette with 3 different levels

Middle age data

middleAgeDate <- cleanedData %>%
  filter(Customer_Age >= 25 & Customer_Age <= 45 & Payment_Method == 'debit card')
  
num_fraudulent_transactions <- sum(middleAgeDate$Is_Fraudulent == 1)
percent <- (num_fraudulent_transactions / num_transactions) * 100
print(percent)
## [1] 3.539624