User Context Analysis from Spatial Interface Interactions

Our work addresses the problem of information overload in the spatial domain. Information overload can confuse and overwhelm users by providing too much detail, often complicating simple interaction tasks. Knowing users’ interests allows the system to alleviate information overload, and therefore facilitate interaction, by tailoring the information displayed at the interface to individual user preferences. Spatial data (e.g. topographic maps) are rich in content and typically require explicit actions such as zooming and panning to view their extent. This paper presents preliminary results which show that user’s spatial interests can viably be ascertained by monitoring mouse movements and clicks as spatial implicit interest indicators. An interest modelling algorithm analyses this interaction information to produce user interest models, reflecting each user’s latest interests as they change over time. These models can be used for the personalisation of user datasets and interfaces, balancing the display content with the relevance of the available information in the system to the user and his context.