Data Analytics Portfolio

Six end-to-end case studies, one per sector

I’m Michael Udousoro, a data analyst based in the UK. Each project below starts from a real business question, uses public data, and ends with a written report, a dashboard, and a recommendation for a named stakeholder, not just a chart.

Every project follows the same pattern: clean SQL, a documented notebook, a one-page case study, a full written report, and where it fits, a live dashboard. The code and full write-ups are on GitHub.

The projects

Transport — Santander Cycles Demand and Rebalancing

Where is Santander Cycles demand most imbalanced across London, and how should bikes be rebalanced? DuckDB analysis of 1.66 million real trips.

SQL (DuckDB) · Python · Tableau

Healthcare — NHS A&E National Trend and Trust Benchmarking

What drives A&E 4-hour target breaches, and which trusts are the clearest outliers? A full year of NHS England’s own provider-level data.

SQL (DuckDB) · Python · Tableau

Retail — Olist Customer Experience, Retention & AI Sentiment

Who are the valuable customers, does late delivery cost you a good review, and what does an AI sentiment layer on real review text add?

SQL · Python · a pretrained NLP model · Tableau

Finance — Credit Risk: Setting a Smarter Approval Cutoff

Can a lender cut default losses without rejecting too many good borrowers? A logistic regression scorecard, priced as a business decision.

SQL (DuckDB) · Python · Excel · Tableau

Marketing — Maven Fuzzy Factory Funnel & A/B Tests

Which channels convert, where does the funnel leak, and did the website A/B tests actually work? A year of SQL “CEO emails” for a growing startup.

SQL · Python · Tableau

Energy — Global Power Plants: the Renewable Transition

How far along is the world’s shift away from fossil fuel generation, and how reliable is the data behind that question? ~28,700 power plants worldwide.

Python · Plotly · matplotlib