Credit Access & Personal Finance Analysis

Business Intelligence · Machine Learning · Data Storytelling

An end-to-end analysis of U.S. credit access using Federal Reserve SCE data (2013–2025), combining exploratory analysis, machine learning models, and policy-relevant insights.

Tech stack

Python · Pandas · Scikit-learn · XGBoost · Optuna · MICE · Matplotlib/Seaborn · Business Intelligence

Problem & Why It Matters

Credit access shapes household financial stability, but access is not evenly distributed. Following the COVID-19 period, concerns grew around rising rejection rates and discouraged borrowers.

This project investigates who gets credit, who gets rejected, and how credit score and age systematically affect access.

Exploratory Analysis

Key variables: Credit access indicator, Debt-to-income ratio, Savings rate, Income level, Demographic controls

Credit cards show the highest approval rates, while auto loans and mortgages experience more volatility, hinting at higher underwriting sensitivity.

Key Findings

Analysis of Federal Reserve Survey of Consumer Expectations (2013–2025) reveals structural patterns in credit access across products, demographics, and credit scores.

Three major insights emerged from the analysis:

1. Credit Score Strongly Predicts Approval Outcomes

Applicants with credit scores above 760 experienced significantly higher approval rates across all debt

products, while applicants with scores below 680 faced rejection rates more than five times higher.

This confirms that credit score remains the most dominant factor in credit accessibility.

2. Credit Card Applications Are More Stable Than Other Credit Products

Credit card applications remained relatively consistent over time compared to mortgage and auto loan

applications, which showed larger fluctuations.

This suggests that unsecured credit products maintain more stable approval pipelines during economic shifts.

3. Lower Credit Score Groups Experience Higher Discouragement

Individuals with credit scores below 680 showed the highest levels of credit discouragement, meaning

they chose not to apply due to expectations of rejection.

This indicates that barriers to credit access extend beyond formal rejection rates.

Business Implication

The analysis highlights several implications for financial institutions and policy makers:

• Credit score thresholds significantly influence lending access
• Mid-tier borrowers may face tightening credit conditions during economic uncertainty
• Credit discouragement suggests unmet financial demand among lower credit groups

Understanding these patterns can help institutions design more inclusive credit products and risk assessment models.

In Conslusion

This project demonstrates my ability to work with raw blockchain transaction data, design meaningful aggregations, and

translate complex stablecoin activity into insights relevant to investors and analysts. Rather than focusing on price, the

analysis examines actual usage behavior, adoption trends, and concentration risks within a decentralized stablecoin

ecosystem.