top of page

Credit Portfolio Analysis: Performance, Risk, and Segmentation

  • Writer: Michael Olaniyi Jeremiah
    Michael Olaniyi Jeremiah
  • 2 days ago
  • 4 min read

An End-to-End Power BI Analytics Solution On Microsoft Fabric For BrightShore Lending



By Michael Olaniyi Jeremiah











The Context

BrightShore Lending is strategically expanding its digital loan operations

across the United Kingdom.


The Current State

This rapid growth has resulted in large volumes of borrower and loan data stored across different fields, creating a lack of a unified view on performance, repayment behavior, and emerging risks.


The Solution

I developed an end-to-end Power Bl Analytics solution to provide a consolidated view, enabling executives to monitor portfolio health, identify risk patterns and make data-driven underwriting decisions.


Business Problem

Despite steady growth in loan applications, BrightShore Lending is experiencing increasing delinquencies and charge-offs. The absence of a consolidated analytics framework makes it difficult to:


Key Pain Points

  • Effectively distinguish good loans from bad loans.

  • Proactively identify high-risk borrower segments.

  • Understand critical performance differences at regional and product levels

  • Detect early warning signals of portfolio deterioration.




A Discipline Approach to Unlocking Portfolio Insights


Data Foundation

  • 37,000+ loan records with 25 fields covering Demographics, Loan Characteristics, and Repayment History.



Data Preparation

Cleaned and standardized inconsistent entries, data types, and key columns like 'loan_status' and 'issue_date'.


Data Modeling

  • I Implemented a star-schema approach with a central 'CreditPortfolio' fact table, a dedicated Date table for time-series analysis, and a custom County dimension for geospatial mapping. (Accompanied by smaller labels: Dimension Table → Dimension Table → Fact Table → Dimension Table → Elaboration Table)




Analysis & Metrics

  • Developed 41 comprehensive set of DAX measures to compute critical KPIs, including YTD, MTD trends, Default Rate, Delinquency Rate, etc.


Measures



























  • I Created 3 measure folders to groups the measures for proper documentation and organization:

    1. Anchor Measures: Measures that summarize data (e.g., SUM, AVERAGE, COUNT, DISTINCTCOUNT). 

    2. Time Intelligence: measures that analyze data over time (e.g., YTD, MTD, SAMEPERIODLASTYEAR).

    3. Variance Measures: Measures that modify filter context to derive insights (e.g., CALCULATE, FILTER, ALL, ALLEXCEPT).


Measure Folders


















Dashboard Wireframe

  • Designed 3 pages of wireframe using Microsoft PowerPoint, the exported as PNG file and imported into power BI services using Microsoft Fabric.

  • Because wireframe ensures the Business Intelligence dashboard is structured around user needs and key business questions before development, reducing rework and ensuring clarity, usability, and alignment with decision-making goals.


Wireframe




Visualization & Reporting

  • Built three interconnected, interactive Power BI Dashboards: Portfolio Overview, Loan Performance, and Borrower Risk.


Portfolio Overview


 Loan Performance



Borrower Risk


The dashboards were designed with tooltip on Monthly trends and map, to give stake holders more drilldown insights.




Insight 1: Steady Growth is Primarily Driven by Debt Consolidation



Key Findings


Controlled Expansion

Lending activity demonstrated steady growth throughout the year, with loan volumes rising from 2,214 in January to 4,095 in December. This indicates a disciplined growth strategy.


Core Purpose

Loan refinancing and credit repayment consistently account for the highest loan volumes, suggesting borrowers are using loans to manage existing financial obligations.


Dominant Demographic

'Middle-Income' earners (£40k-£80k) are the core borrower segment, accounting for 17,629 loans.


Observation:

The portfolio is fundamentally healthy and growing, but its profitability is highly sensitive to the income stability of its core borrowers.




Insight 2: High Returns are Accompanied by Significant Default Risk  





Key Findings

The findings indicate that mid-tier grades, particularly Grades B and C, provide the most balanced contribution to portfolio performance, combining meaningful loan volumes, reasonable yields, and manageable default risk.


While Greater London has the highest loan volume (4,388 loans), Greater Manchester shows a significantly higher relative default rate at 16.3%, requiring closer monitoring.




Insight 3: Affordability is the #1 Predictor of Default


Default Rate Skyrockets as Debt Burden Increases


The Critical Finding Debt-to-Income (DTI) is the strongest driver of credit risk. As a borrower's existing burden rises, the likelihood of default increases significantly,

regardless of other factors.


Implication Current risk assessment may be focused on the wrong indicators. 




Mitigating risk requires accurately assessing a borrower's repayment capacity, rather than focusing solely on identity or income.




Challenging Old Assumptions: What Doesn't Predict Risk


Verification Status is Misleading


  • Verification confirms identity, not affordability.








Employment Length is Not a Safeguard


  • Job stability does not guarantee financial discipline.










Housing Stability is Crucial


  • Housing a powerful proxy for overall financial stability.









Strategic Recommendation: A Data-Driven Path to Sustainable Growth 



1. Prioritize Affordability in All Credit Decisions

  • Place stronger emphasis on DTI thresholds during loan approval.

  • Apply stricter underwriting criteria for borrowers with unstable housing status.



2. Optimize the Portfolio Mix for Risk-Adjusted Returns

  • Focus portfolio growth on the 'sweet spot' of Mid-Tier Grades (B & C).

  • Limit exposure to high-risk Grade G loans.

  • Proactively manage regional exposure in hotspots like Greater Manchester.




3. Refine the Underwriting Process

  • Use Verification as a supporting identity check, not a primary risk filter. It must be combined with DTI and housing stability assessments.



4. Implement an Early Warning System

  • Introduce monitoring that tracks chargeoffs relative to loan growth to catch portfolio deterioration as it scales.




Conclusion: The Path to Sustainable Profitability



Where We Are Today

BrightShore Lending has a healthy and profitable portfolio, generating a £36.68M in profit. The business is built on a foundation of controlled, steady growth.


The Unseen Challenge

However, risk is scaling directly with volume. As the loan book grows, so do delinquencies and charge-offs.

The current underwriting model is not optimized to filter for the primary driver of this risk.


The Strategic Imperative

From: "Identity Verification" ===>> To: "Affordability Analysis"

By prioritizing deep, authoritative terms: Debt-to-Income and Housing Stability.

BrightShore can continue its expansion while protecting long-term portfolio health and profitability.





-----------------------------------------------------------------------------------------------------------------------

Thank you for reviewing this project. I welcome constructive feedback and insights.


For any project discussions or job opportunities, please feel free to contact me at:

Phone: 234706664402

 
 
 

Comments

bottom of page