REGIONAL WEBGIS USER ACCESS PATTERNS BASED ON A WEIGHTED BIPARTITE NETWORK

Abstract. With the rapid development of geographic information services, Web Geographic Information Systems (WebGIS) have become an indispensable part of everyday life; correspondingly, map search engines have become extremely popular with users and WebGIS sites receive a massive volume of requests for access. These WebGIS users and the content accessed have regional characteristics; to understand regional patterns, we mined regional WebGIS user access patterns based on a weighted bipartite network. We first established a weighted bipartite network model for regional user access to a WebGIS. Then, based on the massive user WebGIS access logs, we clustered geographic information accessed and thereby identified hot access areas. Finally we quantitatively analyzed the access interests of regional users and the visitation volume characteristics of regional user access to these hot access areas in terms of user access permeability, user usage rate, and user access viscosity. Our research results show that regional user access to WebGIS is spatially aggregated, and the hot access areas that regional users accessed are associated with specific periods of time. Most regional user contact with hot accessed areas is variable and intermittent but for some users, their access to certain areas is continuous as it is associated with ongoing or recurrent objectives. The weighted bipartite network model for regional user WebGIS access provides a valid analysis method for studying user behaviour in WebGIS and the proposed access pattern exhibits access interest of regional user is spatiotemporal aggregated and presents a heavy-tailed distribution. Understanding user access patterns is good for WebGIS providers and supports better operational decision-making, and helpful for developers when optimizing WebGIS system architecture and deployment, so as to improve the user experience and to expand the popularity of WebGIS.