Neural Network Loan Risk Prediction API

Test this API:

Fill form with a typical

Loan amount

Term

Length of employment

Home onwership

Annual income

Loan purpose

Debt-to-income ratio

Number of delinquencies in the past two years

Age of earliest credit line in months

FICO range (low)

FICO range (high)

Number of inquiries in the last six months

Months since last delinquency Enter a number or "N/A"

Months since last public record Enter a number or "N/A"

Number of open credit lines

Number of derogatory public records

Total balance of revolving credit lines

Revolving line utilization rate

Total number of credit lines in credit file

Number of collections in the last 12 months (excluding medical)

Months since last 90-day or worse rating Enter a number or "N/A"

Application type

Combined annual income of co-borrowers Enter a number or "N/A"

Combined debt-to-income ratio of co-borrowers Enter a number or "N/A"

Number of accounts now delinquent

Total collection amounts ever owed

Total current balance of all accounts

Total revolving credit limit

Number of trades opened in the past 24 months

Average current balance of all accounts

Total open to buy on bankcards

Bankcard credit utilization rate

Number of charge-offs in the last 12 months

Past-due amount owed on which currently delinquent

Months since oldest bank installment account opened

Months since oldest revolving account opened

Months since most recent revolving account opened Enter a number or "N/A"

Months since most recent account opened Enter a number or "N/A"

Number of mortgage accounts

Months since most recent bankcard account opened Enter a number or "N/A"

Months since most recent bankcard delinquency Enter a number or "N/A"

Months since most recent inquiry Enter a number or "N/A"

Months since most recent revolving delinquency Enter a number or "N/A"

Number of accounts ever 120 or more days past due

Number of active bankcard accounts

Number of active revolving trades

Number of satisfactory bankcard accounts

Total number of bankcard accounts

Total number of installment accounts

Number of open revolving accounts

Total number of revolving accounts

Number of revolving trades with balance greater than 0

Number of satisfactory accounts

Number of accounts currently 120 days past due

Number of accounts currently 30 days past due

Number of accounts 90 or more days past due in the past 24 months

Number of accounts opened in the past 12 months

Percent of trades never delinquent

Percentage of all bankcard accounts greater than 75% of limit

Number of public record bankruptcies

Number of tax liens

Total high credit/credit limit

Total credit balance excluding mortgage

Total bankcard credit limit

Total installment high credit/credit limit

API output:

LendingClub is the world's largest peer-to-peer lending platform. Until recently (through the end of 2018), LendingClub published a public dataset of all loans issued since the company's launch in 2007.

I decided to create a neural network with this dataset to predict the fraction of an expected loan return that a prospective borrower will pay back. Now, to share it, I've created this API to serve the model. If you'd like to interface with this API directly (rather than using the above web form), send a GET request to https://tywmick.pythonanywhere.com/api/predict (just make sure to include every parameter in your query string).

This API is written in Python with Flask. I also used TensorFlow/Keras to load and predict with the model I created, and pandas and scikit-learn to manipulate the incoming API data a bit before loading it into the model.