The Uncertainty of Identity: Linking Spatiotemporal Information Between Virtual and Real Worlds

点击翻译按钮获取中文摘要

This is an interdisciplinary proposal from Computer Science (St. Andrews), Engineering (City University London) and Geography (UCL), in partnership with experts in Visual Analytics at Purdue University in the United States. Our goal is to link information pertaining to human characteristics in 'real' and 'virtual' worlds in order to better manage the uncertainties inherent in establishing human identity and linking it to geographic locations. Our basic premise is that uncertainty in identifying and characterising individuals may be managed and understood by: (a) detecting and exploring spatio-temporal profiles of lifestyles and activity patterns; (b) concatenating and conflating detailed, but under-exploited, datasets in the virtual and real domains; and, more speculatively (c) soliciting and analysing crowd-sourced volunteered data that link physical and virtual identities. Through these actions it will be possible to improve our ability to characterize and validate an individual's identity, to infer more informative profiles of individuals and groups that bridge the real and virtual domains, and to document and manage the uncertainties inherent in these tasks. Aspects of this highly innovative research agenda are inevitably risky and speculative, but following an EPSRC 'WDYTYA' Sandpit we have appraised risk, examined the feasibility of data acquisition and addressed ethical approval issues. The research will require multiple sources of data about a user's online activities (henceforth 'virtual' sources, such as multiple social networks, commercial information, purchases, etc.) alongside more conventional data (henceforth 'real' sources: population censuses, names registers, telephone directories, social surveys, etc.). Systematic linkage will be used to better resolve the question "Who do I think you are?" We propose the exploitation of complementary databases and methods in order to relate 'real' and 'virtual' properties, to glean, synergise and cross validate new information and to leverage value from secondary sources. This will be achieved by developing novel methods of data collection, maintenance, exploration, analysis and modelling, that are efficient, effective, scalable, and safe to use.The work programme will be undertaken through a programme of six inter-linked work packages in the UK and US, viz:Work Package 1: Data Collection Tools The development of new and effective tools for virtual data collectionWork Package 2: Text Analytics Development of text analytics algorithms to describe clusters of concepts, or associations between certain concepts or named entities. Work Package 3: Data Anonymisation and Privacy Preservation Achieving a balance between the benefits of enhanced data collection (Work Package 1) and text mining (Work Package 2) versus the imperatives of preserving individual privacy. Work Package 4: Cybergeodemographics Use of primary (Work Packages 1 and 2) and secondary data to relate virtual Internet traffic to the probable physical locations from which it emanated; and the development of typologies of social networks that are robust, generalized and related to physical locations.Work Package 5: Spatio-temporal Network Analysis Development and application of spatio-temporal network analysis techniques to emerging social and geographic networks of individuals and the systems used by them. Work package 6: Visual Analytics Deployment of a range of visual exploratory data analysis techniques to alert users to deviations from trend or average behaviour and profile. http://worldnames.publicprofiler.org/UncertaintyOfIdentity/index.html