At Business Science Corporation, I built a customer-segmentation engine combining k-means clustering and LightGBM classification, then validated its impact with an A/B test. The treatment group showed a 5% increase in usage and spend over the control group.
I also analysed a proposed customer-retention incentive and found its cost would exceed the revenue it was meant to protect. This helped avert a projected R12.4M net loss, and I designed a data-driven alternative to address churn.
For a flagship telecoms engagement, I delivered an analytics suite covering performance reporting, portfolio analysis, segmentation, and churn monitoring. I also designed a Post-Implementation Review framework using cosine-similarity matching to compare campaigns.
Earlier, at Drayton Glendower & Mokhobo, I built a forecasting model that identified market opportunities and contributed to a fourfold increase in successful pitches. I hold an M.Sc. in Data Science and Advanced Analytics from University of Cape Town, where I earned a distinction in coursework.
