A Predictive Data Analytics Methodology for Online Food Delivery

Abstract

Online food delivery (OFD) has become a popular and profitable e-business category due to the rising demand for online food delivery. People are increasingly ordering food online, especially in urban areas and on college campuses. Using data from online food delivery services, one can analyze and predict the values of key performance indicators (KPIs). In the study presented in this paper, we developed a systematic methodology to analyze and predict such KPIs using various classification and regression algorithms. We found that, for the case study we analyzed, Random Forest (RF) consistently ranked as the best algorithm for regression and classification in predicting most of the KPIs. The methodology we introduce and illustrate in the paper can be adapted and extended to similar problems to reveal potential operational issues and identify the possible root causes of such problems.