Forecasting Meal Requirements Using Time Series Methods in Organization

After the Industrial Revolution, organizations are mostly obliged to provide catering services to their employees. The aim of the managers in the organization is customers’ satisfaction on the highest level and no unnecessary loss of food. Forecasting is defined as the prediction of future events based on known past values of relevant variable. The ability to accurately predict expected meal counts allows managers to plan the right amount of food to buy and produce food and to plan appropriate staff levels so that food can be prepared and served efficiently. The purpose of this study is to determine which forecasting model will predict the number of meal counts at university dining facilities in the most accurate way. Forecasting techniques including ARIMA, artificial neural network and Facebook Prophet algorithm are applied to data gathering from dining halls over twelve months. The result of this study is that artificial neural network is the most accurate forecasting method. Facebook Prophet API is another appropriate forecasting method because of its simple use and high-level accuracy. Enriched, accurate and impressive reports are always welcomed by managers. This work will also provide the ability to report forecasts to managers in an understandable, comparable and manageable way.