Problem

Manual spreadsheet planning caused frequent stock-outs on fast-moving SKUs and overstock on slow ones.

Data

3 years of POS transactions, promotion calendar and store attributes (2.4M rows).

Solution

Built a Prophet-based forecasting pipeline with promo regressors and an automated replenishment report per store.

Problem

Churn was rising and the team had no way to know which customers were at risk.

Data

Usage events, billing history and support tickets for 85,000 subscribers.

Solution

Feature-engineered behavioural signals, trained a gradient boosting model and shipped weekly risk scores to CRM.

Problem

Department heads received occupancy and wait-time reports days late, if at all.

Data

Admission, discharge and triage records across 6 departments.

Solution

Modelled a star schema in SQL and built an hourly-refreshed Power BI dashboard with role-level security.