Doing groceries again: towards a recommender system for grocery stores selection

Abstract

Choosing a store (i.e. grocery, restaurant etc.) depends on different decision criteria. If the data for these criteria is distributed among different sources a user might need to invest a substantial amount of time to aggregate the necessary information from different resources or base their decisions only on a subset of criteria. Additionally, visualising all criteria can augment the user's decision making. In this work, we demonstrate a prototype that is able to combine the data of different decision criteria from different (online) resources and provides recommendations of the combined decision criteria. Additionally, a skyline facilitates the choice of stores that dominate specific features. As a concrete example, we state a query of the type "Get me all stores of a supermarket (of a particular company) in the vicinity". The data for the chosen criteria of traffic time, distance, or occupancy of stores were obtained from Google traffic, popular times and timeline. The timeline data is used for our introduced decision criterion 'utility' which is an indicator of how much added value is gained by visiting a particular store. The visualization of this allows users to see at one glance different criteria of their decision-making of which supermarket to choose, which can make the difference between an efficient and hassle-free groceries experience and one that comes at a high cost of money, time and nerves.