Predicting seasonal influenza using supermarket retail records

Ioanna Miliou, Xinyue Xiong, Salvatore Rinzivillo, Qian Zhang, Giulio Rossetti, Fosca Giannotti, Dino Pedreschi, Alessandro Vespignani

Abstract: Author summary Seasonal influenza is a major burden to the health care systems of countries. Machine learning approaches and data from external sources are increasingly used for flu forecasting in recent years. In this study, we explore whether the inclusion of retail records in a predictive model improves seasonal influenza forecasting. Specifically, we consider supermarket retail data as a proxy signal for influenza, through the identification of sentinel baskets, i.e., products bought together by a population of selected customers. We develop a nowcasting and forecasting framework that provides estimates for influenza incidence in Italy up to 4 weeks ahead. Our predictions outperform the baseline approaches thus proving the added value of incorporating retail market data in forecasting models.