Demand-Staffing Optimizer

A demand forecasting and dynamic staffing optimization pipeline for capacity planning, combining time-series forecasting (7-day moving average + Prophet) with a Genetic Algorithm (GA) to produce hire/fire decisions that maximize net profit under real-world constraints: 60-day service backlog, 1-month agent ramp-up, and labor efficiency.
The system guides the user from raw event data through aggregation and MA7 smoothing, Prophet-based demand forecasts (with optional per-region calibration and grid search), and GA-driven staffing plans that balance revenue, salary, hiring/firing costs, and lost revenue from unmet demand—enabling scenario-based capacity planning with a 60-day backlog and 1-month agent ramp.

Authors
Yihan Zhou
(she/her)
Data Scientist
Data Science and Research Fellow at McKinsey & Company. Designs analytics and AI systems that help people make clearer decisions under complexity.