Big Data in Restaurant Management: Unsupervised Modelling of Ticket Data and Environmental Variables for Sales Forecasting

Abstract

Revenue Management (RM) is one of the challenges facing the restaurant industry, mainly due to the lack of technology in this sector and the lack of data. Forecasting is the most valuable input of RM. For this reason, the main objective of this research is the proposal of a sales forecasting model based on the data provided by the tickets of a restaurant to extract information that allows the correct management of price and capacity. A system based on an unsupervised Machine Learning (ML) model was implemented to analyze the information and visualize the relationships between dishes and temperatures. The developed system uses unsupervised ML techniques, such as multicomponent analysis and bootstrap sampling, to identify and visualize statistically relevant relationships between data. This study provides a simple and understandable solution to improve management and maximize profits to support restaurant managers’ decision-making.