UNDERSTANDING THE SOCIAL CONSTITUTION OF PERSONAL DATA. The situated nature of how people may reason about their data and not only its sensitivity but also what that data ‘means’ may actually present some interesting challenges to would-be users of that data, challenges that should not be under-estimated. Whilst concerns about the assembly of digital information in ways that might threaten people’s privacy are understandable, given the range of ways in which information can now be captured, a counter-issue to be borne in mind is the amount of work involved in actually assembling information in that way and then somehow making it intelligible to would-be users. Setting aside for a moment the above observations regarding situated reasoning about sensitivity, the potential usability of things like addresses and bank account numbers and images is relatively evident prior to any stipulation regarding what that use might be. However, the as-is usability of some of the ‘new’ kinds of data about users and their context being postulated in more recent debates, for instance information being gathered by sensors or through the assembly of traces of digital interactions of various kinds, is far less evident. Matters of ambiguity, comprehensibility and the diversity of content are all relevant here. Additionally, other research has emphasized the extent to which much of the data being referred to here requires explication by the people from whom it is gathered before it can make any proper sense. Log data, for instance, is often almost impossible to understand without understanding the exact human context within which it was gathered. To elaborate upon the preceding point, a recent (and as yet unpublished) ethnographic study examined how people managed and shared a range of personal data collected in their homes via a combination of Current Cost energy monitors; motion, humidity, light level and temperature sensors; and a Withings Smart Body Analyser which collated data about household members’ weight, percentage body fat, heart rate, and CO2 levels. The output from these sensors was made available to users through both a simple graph-based visualization that showed peaks and troughs at various times of the day, and the similar graph-based visualization that is provided to all users of Withings devices. A key feature of the study was an emphasis upon examining how people actually reasoned about the data being produced by the sensors as they encountered it. Three key issues were uncovered: the visibility to sensors of the so...
