Personal informatics is a class of tools that help people collect personally relevant information such daily steps taken, weight, blood pressure etc for the purpose of self-reflection and self-monitoring. These include devices like Fitbit, Apple Watch etc.

PI systems provide an advantage over simply trying to remember information about the self, because pure self-reflection is often flawed. People have very limited memory, reflecting by using memory alone makes it difficult to see patterns and trends.

Current State of PI

Activity trackers can be used in five different ways, namely to reach a goal, to document activities, to link different things together, to register achievements, and for a pure interest in data and technology.

The first and most widely-used model describing how people use personal informatics tools is Li et al.’s five stage model of personal informatics systems [1].

• Preparation, where the user starts collecting personal information;
• Collection, where they gathers data about themself;
• Integration, where the information collected are prepared and transformed for the user to reflect upon;
• Reflection, where the user reflects on their data;
• Action, where they chooses how to behave on the basis of their newfound understanding of themself.

This model that informs most PI technologies is borrowed from behavior modification philosophies, resulting in a transfer of emphasis from the self to the behavior to be modified. This establishes a mindset that values actions over the self that causes them.

The self that these tools should help unveil, therefore, actually becomes fragmented in the individual behaviors that have to be modified.

The Problem with current PI applications

Personal Informatics devices simply capture and display behavioral/ physiological results, such as the amount of steps taken or the level of blood pressure. Although their alleged claim is to boost self-reflection and encourage self knowledge, a description of this “self”, to which these knowledge endeavors are addressed, is missing in the current debate within Personal Informatics.

The self that these instruments quantify thus is reduced to the data pattern referred to the single behavior/parameter tracked, and the self-knowledge that they actually provide is mere information about how the user behaved in the past.

How do we fix this?

We need to reframe the ways in which Personal Informatics can produce a change in its users, starting from their inner self instead of their behavior. This needs to be done as our take on reality and ourselves is subjective, and that our knowledge and self-knowledge are dynamically constructed.

Amon et al (2017) [2] further bring up the four different facets of the self: the present self, the past self, the future self and the interconnected self.

• The present self is a flow of meaningful subjective experience.
• The past self is a construction of the present self, a representation of itself at a time which supposedly has already been experienced and which is recalled through the act of remembering.
• The future self consists of representations of ourselves as we might exist in alternate times. The interconnected self consists of all the images of ourselves that come from our social interactions.

The Paradigm Shift

This paradigm of looking at the self in Personal Informatics in the form of four different facets helps us design a better association between the data captured and its meaning to the user.

The Present Self

Our knowledge of the world and of ourselves is intrinsically subjective and is driven by the needs, desires, memories, and expectations of our present self. PI should strive to inform data with personal meanings. For example, PI could invite users to add “subjective tags” to the data supplied so as to reflect their internal states like emotions, intentions, goals, and so on, thus making the subjective relevance of information apparent.

Subjectivity could be better supported by providing representations of data that could make the user closer to the information they collected. for example, presenting connections among behavioral data by using natural language (e.g., “On weeks when you are happier you walk more”).

PI should essentially aim at enhancing users’ self-awareness and self-knowledge, pursuing reflection as an end in itself. This self-understanding will primarily yield a change in the individual’s subjectivity through an enrichment and eventually a reconstruction and reinterpretation of their knowledge of it. In this perspective, an individual may decide what kind of change will benefit themself the most, instead of accepting normative social models, as it is often encouraged by current self-tracking tools.

The Past Self

Our experience of the past is organized around the changes in our identity and the emergence of important aspects of our self. This entails that the activation of our past self-images may lead to experience more vividly a network of memories for specific episodes relating to turning points in our life.

Memories can function as “lessons from the past,” data pertaining to past selves could also serve as resources for the current self’s reflections. PI tools could provide suggestions based on the experiences of the past selves as realtime recommendations that can take into account both the user’s current external (environmental) and internal (e.g. goals, plans, and feelings) contexts. These recommendations may lead to new forms of personalization, flexible and tailored to the individual’s past history as well as to her current situation, which could be used, for example, to provide recommendations on how to interpret their current state and even how to act to meet their goals.

The Future Self

Imagining ourselves in the future is an essential mental tool that prepares us for action. Because we can often anticipate how an event might lead to particular physiological, emotional, and mental states, we can calibrate our present actions according to these predictions.

Having the user think of what they might become in the future may provide her with additional resources for reinterpreting her current situation. Adding predictive features to PI tools could help behavior change by suggesting courses of action and predicting opportunities for positive behavior.

Userscould thus become aware of the consequences of their actions: What will I become if I persevere in my actions? How will I transform myself if I follow an ideal path? How will I decay if I divert from the main road?

The Interconnected Self

The interconnected self consists of all the representations of ourselves that come from our social interactions. We show different selves to the different groups and individuals with which we interact; in turn, each of these social relations contributes to shape our evolving identities.

Comparison with the others, in fact, is a fundamental psychological mechanism that influences people’s judgments, experiences, and identities. The relevance of this for PI is that simply looking at the private self cannot guarantee a satisfactory self-knowledge.

PI systems should exploit this research to design spaces where people can engage in dyadic and group interactions, to disclose their own experiences and embrace those of others, which may lead to a reinterpretation of how they look at themselves.