Project_1_Uyemaa_GantulgaR_1616_1015.rmd

---
title: "Project one: Lending Club Data Analysis"
author: "Uyemaa, Aakash, Melissa, Ayush"
date: "10/15/2024"
output:
  html_document: 
    code_folding: show
    toc: true
    toc_float: true
  pdf_document:
    toc: yes
    toc_depth: 3
---

### Setup chunk: Load libraries and set global options

```{r}
knitr::opts_chunk$set(echo = TRUE, include = TRUE)
library(knitr)
library(dplyr)
library(ggplot2)
library(kableExtra)

```

### Load the accepted and rejected loans dataset

```{r}
accepted_loans <- read.csv("C:\\Users\\uemaa\\Documents\\Data Science\\Project 1\\cleaned_accepted_2013_to_2018Q4.csv")
rejected_loans <- read.csv("C:\\Users\\uemaa\\Documents\\Data Science\\Project 1\\cleaned_rejected_2013_to_2018Q4.csv")
```

### Display column names of accepted loans dataset

```{r}
colnames(accepted_loans)
colnames(rejected_loans)
```

### Structure of accepted loans dataset

```{r}
str(accepted_loans)
```

### Structure of rejected loans dataset

```{r}
str(rejected_loans)
```

### Summary of accepted and rejected loans dataset

```{r}
summary(accepted_loans)
summary(rejected_loans)
```


## Question 3: Do the loan grades provided each customer correlate to their loan repayment behavior based on income and their loan status? 

### **Inspecting the Relevant Variables**

1. Loan Grades ('grade')
2. Income ('annual_inc')
3. Loan status ('loan_status' - )

### Inspecting key columns
```{r inspect_data}
colnames(accepted_loans)
head(accepted_loans)
```


```{r}
table(accepted_loans$loan_status)
table(accepted_loans$grade)

```


### **Filtering the Data Based on Loan Status**
### Let’s focus on loan statuses that reflect repayment behavior, such as "Fully Paid" and "Default."

```{r filter_data}
# Filter for loans that are either fully paid or defaulted
loan_data_filtered <- accepted_loans %>%
  filter(loan_status %in% c("Fully Paid", "Default"))

# Check the distribution of loan statuses
table(loan_data_filtered$loan_status)
```

### ** Analyze Loan Grades vs Loan Status**
### Now we analyze whether loan grades correlate with loan repayment behavior (loan status). 

#### **Visualizing the Relationship Between Loan Grade and Loan Status:**

```{r grade_vs_status_plot}
# Plot the relationship between loan grade and loan status
ggplot(loan_data_filtered, aes(x = grade, fill = loan_status)) +
  geom_bar(position = "fill") +
  ggtitle("Loan Grades vs Loan Status") +
  ylab("Proportion of Loan Status") +
  xlab("Loan Grade") +
  scale_fill_manual(values = c("Fully Paid" = "blue", "Default" = "red"))

```

### **Step 6: Perform a chi-squared test**

#### We can use a chi-squared test to statistically assess whether loan grades (specified as `grade`) and repayment behavior (loan status) are associated.


```{r chi_squared_test}
# Perform a chi-squared test on loan grades and loan status
chisq_test <- chisq.test(table(loan_data_filtered$grade, loan_data_filtered$loan_status))

# Show the results of the chi-squared test
chisq_test
```


### **Step 7: Summarize the Data by Loan Grade**
Finally, we can summarize the average income and repayment behavior for each loan grade.

```{r summary_by_grade}
# Summarize data by loan grade
grade_summary <- loan_data_filtered %>%
  group_by(grade) %>%
  summarise(
    avg_income = mean(annual_inc, na.rm = TRUE),
    fully_paid_ratio = sum(loan_status == "Fully Paid") / n()
  )

# Display the summary in a table
kable(grade_summary, caption = "Average Income and Fully Paid Ratio by Loan Grade") %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"))

```