Evolving the Ecosystem of Personal Behavioral Data

Everyday, people generate lots of personal data. Driven by the increasing use of online services and widespread adoption of smartphones (owned by 68% of U.S. residents; Anderson, 2015), personal data take many forms, including communications (e.g., e-mail, SMS, Facebook), plans and coordination (e.g., calendars, TripIt, to-do lists), entertainment consumption (e.g., YouTube, Spotify, Netflix), finances (e.g., banking, Amazon, eBay), activities (e.g., steps, runs, check-ins), and even health care (e.g., doctor visits, medications, heart rate). Collectively, these data provide a highly detailed description of an individual. Personal data afford the opportunity for many new kinds of applications that might improve people’s lives through deep personalization, tools to manage personal well-being, and services that support identity construction. However, developers currently encounter challenges working with personal data due to its fragmentation across services. This article evaluates the landscape of personal data, including the systemic forces that created current fragmented collections of data and the process required for integrating data from across services into an application. It details challenges the fragmented ecosystem imposes. Finally, it contributes Phenom, an experimental system that addresses these challenges, making it easier to develop applications that access personal data and providing users with greater control over how their data are used.