Peer-to-peer lending
This AE will be completed through code chunks in the lecture slides.
library(tidyverse) Warning: package 'ggplot2' was built under R version 4.5.2Warning: package 'tidyr' was built under R version 4.5.2library(tidymodels)
library(openintro)
library(knitr) Warning: package 'knitr' was built under R version 4.5.2Today’s data is a sample of 50 loans made through a peer-to-peer lending club. The data is in the loan50 data frame in the openintro R package.
We will focus on the following variables:
annual_income_th: Annual income (in $1000s)debt_to_income: Debt-to-income ratio, i.e. the percentage of a borrower’s total debt divided by their total incomeverified_income: Whether borrower’s income source and amount have been verified (Not Verified, Source Verified, Verified)interest_rate: Interest rate for the loanThe goal of this analysis is to use the annual income, debt-to-income ratio, and income verification to understand variability in the interest rate on the loan.
We’ll start with data prep to rescale annual income to $1000’s and recode verified_income to fix an issue with the underlying data.
loan50 <- loan50 |>
mutate(annual_income_th = annual_income / 1000,
verified_income =
case_when(verified_income == "Not Verified" ~ "Not Verified",
verified_income == "Source Verified" ~ "Source Verified",
verified_income == "Verified" ~ "Verified"),
verified_income = as_factor(verified_income)
) glimpse(loan50)Rows: 50
Columns: 19
$ state <fct> NJ, CA, SC, CA, OH, IN, NY, MO, FL, FL, MD, HI…
$ emp_length <dbl> 3, 10, NA, 0, 4, 6, 2, 10, 6, 3, 8, 10, 10, 2,…
$ term <dbl> 60, 36, 36, 36, 60, 36, 36, 36, 60, 60, 36, 36…
$ homeownership <fct> rent, rent, mortgage, rent, mortgage, mortgage…
$ annual_income <dbl> 59000, 60000, 75000, 75000, 254000, 67000, 288…
$ verified_income <fct> Not Verified, Not Verified, Verified, Not Veri…
$ debt_to_income <dbl> 0.55752542, 1.30568333, 1.05628000, 0.57434667…
$ total_credit_limit <int> 95131, 51929, 301373, 59890, 422619, 349825, 1…
$ total_credit_utilized <int> 32894, 78341, 79221, 43076, 60490, 72162, 2872…
$ num_cc_carrying_balance <int> 8, 2, 14, 10, 2, 4, 1, 3, 10, 4, 3, 4, 3, 2, 3…
$ loan_purpose <fct> debt_consolidation, credit_card, debt_consolid…
$ loan_amount <int> 22000, 6000, 25000, 6000, 25000, 6400, 3000, 1…
$ grade <fct> B, B, E, B, B, B, D, A, A, C, D, A, A, A, A, E…
$ interest_rate <dbl> 10.90, 9.92, 26.30, 9.92, 9.43, 9.92, 17.09, 6…
$ public_record_bankrupt <int> 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0…
$ loan_status <fct> Current, Current, Current, Current, Current, C…
$ has_second_income <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALS…
$ total_income <dbl> 59000, 60000, 75000, 75000, 254000, 67000, 288…
$ annual_income_th <dbl> 59.0, 60.0, 75.0, 75.0, 254.0, 67.0, 28.8, 80.…We’ll start by fitting a model in which we include all levels of verified_income.
Fit a model using debt_to_income, annual_income_th, and the indicator variables created below to predict interest_rate.
What do you notice about the model output? Why did this happen?
loan50 <- loan50 |>
mutate(
not_verified = factor(if_else(verified_income == "Not Verified", 1, 0)),
source_verified = factor(if_else(verified_income == "Source Verified", 1, 0)),
verified = factor(if_else(verified_income == "Verified", 1, 0))
)# add code hereNow let’s take a look at the design matrix for the model with predictors debt_to_income, annual_income_th, and verified_income.
How does R choose the baseline level by default?
## add code hereWhat is the intercept for individuals with
Not verified income?
Source verified income?
Verified income?
Fit the model with the predictors debt_to_income, annual_income_th, verified_income , and the interaction between annual_income_th and verified_income.
Neatly display the model results using 3 digits.
# add code hereWrite the estimated regression equation for the people with Not Verified income.
Write the estimated regression equation for people with Verified income.
In general, how do
indicators for categorical predictors impact the model equation?
interaction terms impact the model equation?
Sometimes, you will need to include mathematical notation in your document. There are two ways you can display mathematics in your document:
Inline: Your mathematics will display within the line of text.
Use $ to start and end your LaTex syntax. You can also use the menu: Insert -> LaTex Math -> Inline Math.
Example: The text The simple linear regression model is $\mathbf{y} = \mathbf{X}\boldsymbol{\beta} + \boldsymbol{\epsilon}$ produces
The simple linear regression model is \(\mathbf{y} = \mathbf{X}\boldsymbol{\beta} + \boldsymbol{\epsilon}\)
Displayed: Your mathematics will display outside the line of text
Use a $$ to start and end your LaTex syntax. You can also use the menu: Insert -> LaTex Math -> Display Math.
Example: The text The estimated regression equation is $$\hat{\mathbf{y}} = \mathbf{X}\hat{\boldsymbol{\beta}}$$produces
The estimated regression equation is
\[ \hat{\mathbf{y}} = \mathbf{X}\hat{\boldsymbol{\beta}} \]
Click here for a quick reference of LaTex code.
To submit the AE: