Hotmap: Looking at Geographic Attention

Understanding how people use onlin of the map that are most seen by users. Online the most-accessed URLs. Hotmap takes adva heatmap of the log files over the original maps. discusses the imagery acquisition task that moti visible. We discuss the design choices behind H and tuning images to explore both infrequently-v Index Terms—Geographical visualization, GIS, 1 INTRODUCTION Online interactive mapping systems are growing Microsoft, Google, and Yahoo all offer interactiv varying features and qualities of aerial photogr maintainers and designers of such systems, the behav provide critical cues on how both to improve their understand how users are now interacting with the sys The tools needed to track user data for these system from the traditional tools of log file analysis. M advantage of being tied to common-sense geography In contrast, many log-file analysis tools must create an in which to situate a node-link diagram representi [9][15]. Mapping systems lack hyperlinks: while systems logically lead users between pages, online scrolled and navigated without clicking links. Appl analysis tool to these systems fails to represen dimensions of a mapping system: reconstructing pages by examining referrer logs, for example, ca simple reporting of the most-requested page requ interpretation: presenting the filenames for individual far from intuitive. Instead, the analysts’ understandin can be harnessed to build meaningful log-file vis showing the logs as a map. 1.1 Harnessing Users’ Sense of the World A map of users’ interaction with a space provides inf how they understand and interact with the world. W they find most significant? Where do they look at map where do they look at aerial imagery? How do they the space of the map? There is a long tradition of research studying how p and interact with the physical spaces around them. Ke interviewed people within three cities, asking them to the city that they were familiar with. He found th highly “imageable”—people clearly knew where they neighborhoods were oriented—while other cities, Angeles, were far harder to understand. (Even within areas were better defined than others). William • Danyel Fisher is with the VIBE Lab at Microsoft R danyelf@microsoft.com. Manuscript received 31 March 2007; accepted 1 August 20 27 October 2007. For information on obtaining reprints of this article, pleas tvcg@computer.org. king at Geographic Attention Danyel Fisher ne maps allows data acquisition teams to concentrate their efforts o maps represent vast databases, and so it is insufficient to simply antage of the design of a mapping system’s imagery pyramid to Users’ behavior within the system can be observed and interpret vated Hotmap, and presents several examples of information that H Hotmap, including logarithmic color schemes; low-saturation backg viewed and frequently-viewed spaces.