A Study on the Relationship between the Number of Visitors to Stores Measured by GPS Location Information and Weather

Abstract

Recently, the demand for visitor prediction has been growing especially in retailing industry. Accurate predictions improve efficiency and sales in stores by optimizing selection of products and redusing stock. In this study, we show that weather data is useful for predicting visits using GPS data obtained from smartphones and weather data obtained from web api. For this purpose, we utilized a machine learning model for estimating the number of visitors to each store from weather data, and compared it with the case without weather data. We also evaluate the meteorological conditions caused by visits by analyzing the created model.