A Visual Analytics Approach to Exploration of Hotels in Overlaid Drive-Time Polygons of Attractions Chong Zhang 1,2(B ) , Zhengcong Yin 3 , Peng Gao 2 , and Sathya Prasad 2 1 University of North Carolina at Charlotte, Charlotte, NC 28262, USA 2 Environmental Systems Research Institute, Inc., Redlands, CA 92373, USA czhang@esri.com 3 Texas A&M University, College Station, TX 77843, USA Abstract. Given multiple attractions to go to, choosing hotels is a Bichromatic Reverser k Nearest Neighbor searching problem. Many efforts have been made to address the problem in the domain of spa- tial databases. However, most of them suffer from the difficulty of visual selection and comparison. To integrate human intuition into the analysis process and facilitate information communication with users, geospatial visualization techniques have been developed. Unfortunately, too much information on the map, such as hotels and attractions, and drive-time polygons, could overwhelm users. Following the focus+context principle of information visualization, we propose a visual analytics approach to provide users with the flexibility of making their hotel choice. It consists of three main components: an overlap-free and space-efficient clustering algorithm to reduce the searching space for attractions in the map view (MapView), visual comparison of attraction-hotel drive-time in brush- able Small Multiples View (SMView), and the coordination between the two views. We demonstrate the utility of this approach and explain how MapView and SMView can help address the problem. Keywords: Reverse nearest neighbor · Visual analytics · Spatial clustering · Overlap-free 1 Introduction While choosing a hotel is a multi-criteria decision (cost, rating, class, facility, transportation, etc.) at a conceptual level, the problem based on spatial proxim- ity is still interesting. The hotels selected based on spatial proximity are often considered as a superset of the final one that user will choose to stay. It’s not a easy task. If the preference of these attractions are equally considered, the ideal hotel location candidates should be near the “center” of the polygon connect- ing each attraction such that they have the same and shortest distance to the Live demo of the interface is available here. This work is part of Chong Zhang’s internship at ESRI. c  Springer Nature Switzerland AG 2019 Y. Kawai et al. (Eds.): W2GIS 2019, LNCS 11474, pp. 28–40, 2019. https://doi.org/10.1007/978-3-030-17246-6_3