Back to the work
Finance & research / Analytical study · synthetic data

Loan Application Credit Risk

Look at the combination, not one indicator.

SQL and Power BI analysis of 45,000 synthetic loan applications, with risk segmentation and review-process recommendations.

Finance & research

Understand the signals behind the segment.

01Prepare applications
02Compare outcomes
03Segment risk
04Propose review tiers
01

The question

How do borrower and loan characteristics relate to repayment outcomes in a synthetic dataset? The project examines combinations of characteristics rather than relying on a single indicator.

02

Analysis

SQL cleans, segments, and aggregates 45,000 synthetic application records. Power BI views examine repayment status, default rate, loan amounts, interest rates, credit-score bands, and loan purposes.

03

Process recommendations

The study proposes standard review, enhanced verification, and risk-based escalation for combinations of stronger risk indicators. KPI definitions, dataset assumptions, segmentation rules, and workflow recommendations are documented.

04

Scope

This is an analytical portfolio study on synthetic records. It was not deployed to decide real applicants’ eligibility, and the observed relationships should not be read as a validated production lending policy.

KEEP EXPLORINGLogistics Delay Prediction