projects
Each project gets a full write-up: the question, the data, the approach, the results, and the limitations. Statuses are kept honest as work progresses.
portfolio in progress
Built in public, projects marked “planned” show the question and intended approach now, and get complete write-ups with results as they’re finished.
- 2026-07-20Complete
Customer Churn Prediction
A churn model framed around a real decision (a retention team that can only call about 200 customers) and evaluated against the heuristics it has to beat, not just on accuracy.
classification · scikit-learn · model evaluation · statistics
- 2026-06-20Complete
Bank Term Deposit Predictor
A deployed classifier that predicts whether a client will subscribe to a bank term deposit, with a live Streamlit app for real-time predictions.
classification · random forest · streamlit · deployment
- 2026-07-24Complete
NBA Game Prediction: Can Features Beat Elo?
Predicting NBA winners over 68 years of games, benchmarked honestly against FiveThirtyEight's Elo. A lesson in how hard a good baseline is to beat.
classification · temporal validation · calibration · sports analytics
- 2026-07-18Complete
Revenue & Retention Analytics
An advanced-SQL investigation into why a subscription business's revenue growth was slowing while signups hit records, ending in a campaign A/B test.
sql · window functions · cohort analysis · a/b testing
- 2026-07-22Complete
Retail Operations Dashboard
From a million messy real invoice lines to a four-view decision dashboard, with every cleaning choice counted and every KPI defined and tested.
dashboards · streamlit · data cleaning · kpi design
- 2026-07-25In progress
Demand Forecasting: 7 Days Ahead
A 7-day-ahead demand forecast done the way time series must be: walk-forward backtesting, baselines first, and strict control over what's known at forecast time.
time series · forecasting · backtesting · scikit-learn
- 2026-09-01Planned
Customer Segmentation & Uplift
A planned next project: move beyond ranking who will churn to grouping customers by behaviour and estimating who a retention offer would actually change.
clustering · rfm · uplift modeling · scikit-learn