Selecting a proper warehouse location serving to satisfy the demands of the goods from a certain business area is important to a successful retail business. However, the large solution space, uncertain traffic conditions, and varying business preferences impose great challenges on warehouse location selection. Conventional approaches mainly summarize relevant evaluation criteria and compile them into an analysis report to facilitate rapid data absorption but fail to support a comprehensive and joint decision‐making process in warehouse location selection. In this paper, we propose a visual analytics approach to facilitating warehouse location selection. We first visually centralize relevant information of warehouses and adapts a widely‐used methodology to efficiently rank warehouse candidates. We then design a delivering estimation model based on massive logistics trajectories to resolve the uncertainty issue of traffic conditions of warehouses. Based on these techniques, an interactive framework is proposed to generate and explore the candidate warehouses. We conduct a case study and a within‐subject study with baseline systems to assess the efficacy of our system. Experts ‘feedback also suggests that our approach indeed helps them better tackle the problem of finding an ideal warehouse in the field of retail logistics management.

If it is the author's pre-published version, changes introduced as a result of publishing processes such as copy-editing and formatting may not be reflected in this document. For a definitive version of this work, please refer to the published version.

Figure 1: ShuttleVis includes (A) a dataset loader and data description; (B) overview of car-hailing reimbursement records across different departments and descriptions of the departure and arrival time; (C) directional clustering configuration view to help analysts identify appropriate travel directions;(D) map view to visualize identified directional and regional clustering results, and comparative ranking view that includes (E1) a ranking of shuttle bus stops in terms of (E2) metrics in consecutive regional clusters along one travel direction, (E3) timetables of selected shuttle routes, and (E4) radar chart showing attribute distributions of selected routes.