The MyRoR platform was designed to collect and provide context information that is about and relevant to user’s lifestyle. In order to address real-life scenarios and capture as many aspects as possible of our daily lives without the need for intrusive sensors, I decided to focus on three main categories of sources: (1) wearable sensors; (2) mobile phones; and (3) desktop/laptop machines. The reason for focusing on computing machines and mobile phones is their pervasiveness as well as the huge amount of information they can provide about what we do or think during the day. The wearable sensors were necessary to be included as the only way to automatically collect physiological data necessary to create correlations between events that happened and physiological states, such as stress. The platform can collect a large variety of information from these sources, as shown in the figure. The information is usually collected on these sources, then collected and stored into a MySQL database.
Given the high level of information that can collected as well as derived, I chose a model where the data is stored and processed on a user-side machine. Even though the current trend seems to be that more and more “cloud” services collect user-related information either for adapting functionalities or delivering customised information, my aim is to design and build systems that put information safety and user privacy at the core, by storing the information under full user control over access, usage and sharing. I consider such model to be much more suitable for lifestyle management systems, considering the level of sensitivity associated to the captured data. In the same time, I did not want to create a closed system, therefore connections to external entities are also addressed either through obtaining certain relevant content from web services (e.g. weather providers, Google maps), through accessing remotely stored user data (such as emails and calendar events) or through various sharing options (either of individual types of data or correlated information).
The various types of information processing available in MyRoR are discussed here.