As homes become more capable to infer and interpret human behaviours, so too does the ability of care providers and consumers to improve the conditions of domestic occupancy. Smart-home environments can provide foundational data (e.g. spatial movement), and interpreted behavioural analytics (e.g. food preparation patterns, Brennan, et al. 2013), which later inform decisions by professionals and consumers to improve infrastructure, appliances, and the physical and social environment (e.g. Maguire, et al. 2014). Implicit in this process is the assumption that there is a smooth transition from sensed data to sensible decisions. While post-capture data-fusion and pattern-recognition provide context for professional decision-making, it is less clear how or why end-users would make sense of and operate on collected data.
This project will take a user-centred design approach to co-produce with stakeholders, including end-users, the design of visual interfaces for everyday data management. The use of commercial in-home sensing (e.g. security sensors) can be understood as unproblematic by end-users (Stringer et al. 2006), both in their presence and their validity. However, the value of sensing devices (e.g. fire alarms) may be undermined by the interpretation of its output (e.g. false alarms as positive), or indeed end-user configuration. Consequently, there is a need to develop interfaces for visualising data to support end-users’ sense-making of both sensed data and the function of sensors. Moreover, in exploring the deployment of telecare systems, Vines et al. (2013) have highlighted the sensitive nature of in-home sensed data, particularly when it can be shared with others. This suggests that when collecting large-scale in-home sensing data, there is a significant need for user-facing data governance tools that allow for the obfuscation or removal of personally sensitive data, and the annotation of ambiguous data. Furthermore, the persistence of digital datasets and the increased desire and requirement for open datasets, require new forms of stakeholder participation and consent (Cummings, et al. 2015). Therefore, while increasing the legibility and ownership of datasets by end-users, there is a simultaneous need to account for the provenance of data and allow for greater control over where and why personal data is used.
Funding: Engineering and Physical Sciences Research Council – SPHERE Partnership Funding
Collaborators: Rob Comber (Newcastle University, Principal Investigator)
Timescale: November 2015 to October 2016