21/08/2026
I almost did not post this project.
I completed it months ago, but kept telling myself, “It can still be better.” So I kept polishing instead of putting the work out there.
Eventually, I realised something important: visibility is part of growth.
This is the first level of a deeper Credit Analysis project, built with dummy data to protect confidentiality, especially as a newly Certified Data Protection Officer.
🎯 Objective
To understand the health of a loan portfolio, identify the level of risk, locate where the risk is concentrated, and determine which segments require closer attention.
Using 5,000 loan records, I analysed Product Type, Channel, State, Loan Amount, DTI Ratio, Credit Score and Past Due Days.
📊 Key Findings
• ₦8.2B remains outstanding from ₦12.64B originally disbursed.
• PAR30 stands at 19.16%, indicating significant early-stage delinquency.
•PAR90 is 3.24%, suggesting that a large portion of delinquent loans is still being recovered before reaching the NPL stage.
• BNPL recorded the highest PAR30 at 20.6%, while **Asset Finance had the highest NPL rate at 3.95%.
This revealed two different risk areas. Looking at only one metric could easily hide the bigger picture.
I also found patterns across credit score bands that raised questions about how reliably the score is predicting actual portfolio risk.
And that, for me, is where real analytics begins.
Not just finding numbers, but questioning them, finding patterns and asking better business questions.
Level 2 of this analysis is coming.
I almost kept this project hidden. I'm glad I didn't.
