E-shopping and walking accessibility to retail

Abstract

Previous research on e-shopping and transportation has paid limited attention to accessibility-oriented approaches. To address this gap, the present paper explores the importance of e-shopping habits in predicting three types of walking accessibility (real, potential, and perceived) to grocery and clothing stores for individuals younger than 55 years. The municipality of Getafe (Madrid Metropolitan Area, Spain) provides the empirical staging ground. The methodology utilizes a face-to-face questionnaire to record e-shopping habits and socio-economic characteristics. The Random Forest algorithm is used to establish the importance of e-shopping habits and socio-economic characteristics for predicting real, potential, and perceived walking accessibility. The results reveal that both e-shopping habits as well as socio-economic variables are relevant for predicting each type of walking accessibility, but with varying impacts and targeted in specific use cases. For groceries, e-shopping habits demonstrated high predictive power for potential accessibility, whereas such predictive power was similar to socio-economics for real and perceived accessibility. Regarding clothes, e-shopping habits were more relevant than socio-economics in predicting real and potential accessibility, while a similar importance level was seen between e-shopping habits and socio-economic factors for perceived accessibility. The concluding remarks provide a summary of the key findings and a few reflections on transport policymaking.