This website and Credit Approval App were created as the final project for the Vanderbilt University Data
Analytics Boot Camp. Our Project Team collaborated on a Neural Network (NN) Model with the objective to read
in credit data and create a model to predict an outcome of Credit Risk for potential applicants.
Our NN model was based on previous customer Default patterns from our selected dataset, The
Credit_Data_Original.csv. It contained individual prior histories with 30 features such as: loan detail
categories, purpose of loan, financial information, and personal information such as employment and years of
residency. In order to produce the Default model and analyze outcomes, we employed Jupiter Notebook to import
key libraries: Pandas, Matplotlib.pyplot, Numpy, SkLearn, and Tensorflow.keras.
Utilizing Random Forest, Feature Importance was sorted on our Credit_Data_Original.csv. This allowed the final
Credit_Data_Revised.csv to be narrowed to the top ten features of importance to optimize results when training
the NN model.
The Top 10 Features were derived from the questions on the Application Homepage. These features were then used
in the Default Model to run an applicant’s input and return an approval status.
Top 10 Features of Importance: Loan Amount, Checking Acct Bal, Applicant Age, Loan Duration,
Credit History, Years Employed, Savings Acct Bal, Install Rate, Years of Residency, Job Type
Software Tools Used: Python Pandas, Python Matplotlib, Azure Container Apps, Git, Flask,
SkLearn, R Studio,
Tableau, HTML/CSS/Bootstrap, ULEAD, Snagit, Microsoft Office 365
Website Logo Photo: Stock photo ID:90382959
https://www.istockphoto.com/photo/nashville-skyline-gm90382959-4159728 License purchased: 04/28/20