Featured Project: Customer Segmentation and Loan Prediction with SAS Viya
An in-depth look at a project where I utilized SAS Viya for advanced customer segmentation and loan prediction modeling.
Objective
- Develop customer segments based on behavior and demographics.
- Build a predictive model for loan defaults using SAS Viya's advanced analytics capabilities.
- Segment banking customers into distinct groups and build models to predict loan approval likelihood.
Approach
- Employed SAS Viya's visual analytics to explore customer data and identify key segments.
- Performed exploratory data analysis on income, transactions, credit scores, and loan status.
- Applied K-Means Clustering to divide customers into 10 meaningful segments:contentReference[oaicite:0]{index=0}.
- Built Logistic Regression and Decision Tree models to predict loan approvals.
- Compared models on Accuracy, Precision, and F1 Score.
Results
- Identified five distinct customer segments, enabling targeted marketing strategies. The loan prediction model achieved a 90% accuracy rate, significantly improving risk assessment.
- Identified 10 customer groups: High-Income Savers, Young Professionals, Retirees, etc.
- Logistic Regression achieved 71% accuracy, while Decision Tree achieved 74% precision.
- Recommended Decision Tree model for minimizing false approvals.
Project Files
Banking Customer Dataset (Excel)
This Excel file contains a fictional banking customer dataset with variables such as Customer ID, Age, Gender, Annual Income, Account Balance, Number of Transactions, Total Transaction Amount, Loan Status, and Credit Score. It provides the foundation for data mining tasks like EDA, clustering, and loan approval prediction models in SAS Viya.
👉 Download the dataset here to explore the raw data used for analysis: https://www.dropbox.com/scl/fi/cpt05wobftmasvkxe918r/banking_customer_data-1.xlsx?rlkey=g5378apcnlmmlicjb6vfdp5c4&st=pylli1qp&dl=0
SAS Viya Banking Analysis Report PDF
This PDF showcases a full banking customer data mining project in SAS Viya, featuring detailed exploratory analysis of demographics, balances, incomes, and credit scores, 10-cluster segmentation with tailored marketing strategies, and loan approval prediction models (Logistic Regression vs. Decision Tree). The analysis highlights key drivers—credit score, income, and transactions—and recommends Decision Trees for minimizing false approvals
How to View and Understand This Project?
To fully explore this project, follow these steps:
Start with the Excel Dataset
- Download the dataset or click “View Excel File”.
- Open it in Excel (or Google Sheets) to see the raw customer data, including demographics, balances, transactions, loan status, and credit scores.
- This dataset is the foundation for all subsequent analysis.
Review the PDF Report - Download the SAS Viya Analysis PDF.
- The report explains how the dataset was analyzed using SAS Viya through three steps:
- Exploratory Data Analysis (EDA): Understanding trends and correlations.
- Clustering: Grouping customers into 10 unique segments with marketing/loan strategies.
- Classification Models: Comparing Logistic Regression and Decision Tree for predicting loan approvals.
Interpret the Insights - Look at the EDA results to understand the overall customer profile.
- Review the cluster analysis to see how customers are segmented.
- Compare the classification models to understand why Decision Trees are more reliable for reducing false loan approvals.
By following these steps, you’ll gain a clear understanding of both the raw dataset and the data mining process that transforms it into actionable banking insights.