Integration and Exploration of Connected Personal Digital Traces

Abstract

A large number of personal digital traces is constantly generated or available online from a variety of sources, such as social media, calendars, purchase history, etc. These personal data traces are fragmented and highly heterogeneous, raising the need for an integrated view of the user's activities. Prior research in Personal Information Management focused mostly on creating a static model of the world (objects and their relationships). We argue that a dynamic world view is also helpful for making sense of collections of related personal documents, and propose a partial solution based on scripts -- a theoretically well-founded idea in AI and Cognitive Science. Scripts are stereotypical hierarchical plans for everyday activities, involving interactions between mostly social agents. We augment these with hints of the digital traces that they can leave. By connecting Personal Digital Traces through scripts, we can build an episodic view of users' digital memories, which allow users to explore related events and actions in an integrated way. The paper uses the Eating_Out script for illustration, and ends with a report on the results of a case-study of applying a prototype implementation on real user data.