I help e-commerce and retail businesses turn their data into decisions that generate measurable revenue. Based in Nigeria, working remotely worldwide.
Most e-commerce and retail businesses aren't short on data — they're short on the clarity that data should produce. I became a data scientist because I kept watching businesses sit on gold mines of customer and transaction data, making gut-feel decisions that cost them revenue, customers, and growth.
Over the past five years, I've built machine learning systems that predict customer churn 60 days before it happens, identified $500K in annual cart abandonment losses for online stores, and built Power BI dashboards that leadership teams open every single morning. The common thread isn't the technology — it's that every project starts with a business question, not a model choice.
My work spans the full data science stack: from scraping and cleaning raw data, to building and deploying predictive models, to visualizing insights in interactive dashboards. I've worked across industries — e-commerce, retail, fintech, real estate, and EdTech — and I've learned that the fundamentals are always the same: understand the business, understand the data, and build something that changes a decision.
Beyond client work, I build for domains I care about. My Yoruba-English translation model is a small contribution to the growing ecosystem of African language NLP. My Nigerian used car pricing model exists because most Nigerians buy used cars but have no reliable pricing reference. My Chicken Republic location analysis started as curiosity about how a Nigerian brand expands — and became one of the most-read pieces I've written.
I'm based in Nigeria and work remotely with clients worldwide. If you have data and a revenue problem you can't solve, I'd like to hear about it.
Four deep areas of expertise built over five years of e-commerce, retail, and fintech analytics work.
Building dashboards that connect messy multi-source data into a single view your team actually opens every day. DAX, Power Query, data modeling, star schemas.
Production machine learning models — churn prediction, lead scoring, price prediction, demand forecasting — that run in real environments, not just Jupyter notebooks.
Multi-label classification, sentiment analysis, RAG systems, and language model fine-tuning — including low-resource African language NLP.
RFM analysis, cohort analysis, customer lifetime value modeling, and behavioral segmentation — turning transaction history into strategic customer intelligence.
The model choice is always secondary. Understanding what decision the analysis needs to support, and for whom, determines everything else. I spend more time on scoping than on coding.
A deployed model with 80% accuracy that runs in production generates more value than a 95% accuracy model living in a Jupyter notebook. Iteration beats perfection every time.
If a business team can't understand why the model is recommending what it recommends, they won't trust it — and they'll be right not to. Explainability isn't a nice-to-have, it's a requirement.
Every project I deliver includes documentation so thorough that you never need to come back to me unless you want to. Self-sufficiency for the client is a success metric, not a threat to my business.
I believe African languages deserve the same quality of language technology as English and French. My Yoruba-English translation model is part of a broader personal commitment to building for underrepresented languages.
I publish data-driven analyses on Medium — from the Chicken Republic expansion strategy to customer retention frameworks. Writing forces clarity of thinking that makes the data science better.
Nigeria is one of Africa's most complex and most under-analyzed markets. I build projects specifically for Nigerian data — used car pricing, restaurant expansion, fintech — because local market intelligence matters.
If you have a data problem in your e-commerce or retail business, I'd like to hear about it. First conversation is free — no pitch, no obligation.