A KPI moved. Nobody knows why.
Revenue, margin, volume, price, mix, segment, and channel analysis to isolate what changed, quantify the drivers, and identify where management should investigate next.
PROBLEMS I SOLVE
Bring the business problem. I build the evidence.
The fastest way to evaluate the portfolio is to match your problem to the proof below. Tools are supporting methods, not the product.
The numbers do not reconcile.
Data-quality checks, KPI definitions, reconciliation, analytical SQL, reporting layers, and BI specifications that make metrics auditable before they reach a decision-maker.
Management needs to compare scenarios.
Forecasts, downside/upside cases, break-even, unit economics, ROI, and sensitivity analysis for decisions involving price, volume, cost, cash, and growth assumptions.
An analytical output needs a reliability check.
Risk segmentation, model validation, evidence traceability, AI evaluation, and QA controls that distinguish a usable analytical output from an output that only looks plausible.
BEST-FIT MANDATES
Revenue / KPI diagnosisSQL + BI reportingFinancial & quantitative modellingData quality & analytics QAResearch & AI evaluation
SELECTED WORK
Case studies first. Code second.
The website carries the business story and evidence. GitHub remains the technical appendix for source code, data lineage, tests, outputs, and reproducibility.
AI-POWERED RETAIL SALES DIAGNOSTIC
Why did revenue fall despite increasing sales volume?
A full decision diagnostic from raw-data quality controls to KPI construction, price-volume decomposition, mix analysis, anomaly triage, and recommendation design.
WHAT THIS PROVESI can move from “revenue is down” to a quantified driver narrative without overstating causality.
−14.8%revenue H1→H2
+48.3%units H1→H2
−33.1%ASP H1→H2
−23.4%profit H1→H2
EXECUTIVE DIAGNOSIS
Volume increased, but realized value per unit deteriorated.
The synthetic case points to value dilution: lower-value mix and heavier discounting coincided with weaker revenue, profit, and margin outcomes. The evidence supports diagnosis, not causal attribution.
Retail SQL & BI Analytics Layer
Problem solved: Turn a raw retail dataset into a reliable SQL-backed BI reporting layer that management can interrogate.
Proves KPI design, analytical SQL, dimensional comparisons, and a decision-ready BI structure.
Validated KPI views, analytical SQL, management comparisons, dimensional analysis, and a Power BI build specification that extend the flagship case.
6 core KPI measures4 business dimensionsH1/H2 comparison layer
Nigeria Development Data Monitor
Problem solved: Turn official public data into a compact, auditable monitor for economic and development context.
Proves public-data sourcing, metric definition, comparison, and decision-oriented communication.
Uses World Bank indicators to demonstrate applied data sourcing, metric definition, comparison, and decision-oriented communication outside synthetic datasets.
23.0% 2025 inflation237.5m 2025 population62.5% 2024 electricity access
Credit Risk Analytics & FICO Segmentation
Problem solved: Translate borrower characteristics into measurable risk segments that can support consistent credit decisions.
Proves model evaluation, segmentation logic, and quantitative risk communication.
50,000 borrowers0.70 test AUC8 FICO buckets
Financial Planning & Scenario Modelling
Problem solved: Show how operating assumptions change revenue, profit, cash generation, and downside/upside outcomes.
Proves forecast structure, scenario analysis, sensitivity testing, and model QA.
3 scenario casesQA model checksSensitivity analysis
Pricing, Unit Economics & ROI Engine
Problem solved: Quantify the commercial effect of changing price, volume, and cost assumptions before a decision is made.
Proves unit economics, break-even logic, ROI scenarios, and sensitivity analysis.
Break-even engineROI scenariosSensitivity matrix
Data Quality & Analytics Assurance
Problem solved: Test whether a dashboard or KPI is trustworthy enough to support a decision.
Proves schema validation, reconciliation, QA rules, and explicit PASS/REVIEW/FAIL controls.
Schema validationKPI reconciliationPASS/REVIEW/FAIL
AI Research & Evaluation Framework
Problem solved: Measure whether AI-assisted research is accurate, complete, traceable, and economical enough to rely on.
Proves evidence checks, evaluation criteria, traceability, and cost/latency awareness.
Accuracy & completenessTraceabilityCost/latency
PROOF
Claims map to inspectable evidence.
Synthetic projects demonstrate method. Public-data work demonstrates applied sourcing. Client impact is only claimed when directly evidenced.
Can I inspect the work?
Yes. Source code, data-generation logic, assumptions, validation rules, tests, outputs, and methodological limitations are available for inspection.
Review technical evidence ↗Can I understand it quickly?
Yes. The browser-first case-study layer puts the problem, method, evidence, conclusion, and limitations ahead of the technical appendix.
Browse decision evidence ↗What role does this evidence support?
Data & Quantitative Analyst is the core profile, with BI, financial analysis, research, and AI evaluation presented as connected analytical capabilities.
Read the CV ↗HOW I WORK
Make the reasoning inspectable.
The differentiator is a disciplined path from question to recommendation, with assumptions, controls, and uncertainty visible.
- 01Frame
Define the decision, metric, grain, audience, and constraints.
- 02Inspect
Understand the source, schema, coverage, provenance, and definitions.
- 03Validate
Test quality, assumptions, reconciliations, identities, and edge cases.
- 04Analyse
Quantify drivers, patterns, scenarios, uncertainty, and alternatives.
- 05Explain
Separate observed facts from interpretation and causal claims.
- 06Recommend
Connect the evidence to practical next actions, tests, and monitoring.
NEXT STEP
Need analysis that moves a decision forward?
I’m open to remote Data Analyst, BI Analyst, Quantitative Analyst, research, and analytics project opportunities. Send the decision, data context, or role brief—not just a tool list.