A Data-Driven Dashboard for Karma Kebab Restaurant
In today’s gastronomy industry, operational efficiency and data-driven decision-making are critical for profitability. Many businesses struggle with staffing optimization, demand forecasting, and leveraging operational data effectively.
The project addressed profitability challenges at Karma Kebab caused by overstaffing, understaffing, and limited operational insight. A data-driven dashboard was developed to provide optimized workforce schedules and support more informed decision-making.
The students built a Random Forest model to predict hourly order volume one day ahead, achieving 87% accuracy, and used machine learning and time series models to forecast daily revenue, with a mean absolute error of €639. Additionally, a Two-Stage Stochastic Integer Programming (TSSIP) model was implemented to optimise workforce scheduling, reducing weekly labor costs by 25% compared to a deterministic baseline.