Removing Inferences at The Interface

Wearable devices, health trackers, and sport fitness equipment have become ubiquitous accessories in our day to day. They offer users and wearers the ability to qualify and quantify themselves in ways that used to be inconceivable and would otherwise be unachievable without the technology. Simultaneously, they support better performance while augmenting previously reliable systems.

As they currently stand, personal physiological visualization methods, principally recovery scores, heart rate, encourage behavior change, enhance social presence, create emotional connection, and increase empathy between collaborators or friends. However, to do so, current interfaces have to maintain curiosity of exploration which usually commands a high attentional demand, high explorability, but very low actionability, especially for novice users. These characteristics lead to inconsistencies in how data is presented across companies, across platforms, and across wearables. For instance, there is an idiosyncratic blue choice for exercises in Zone 1 on the 5-Zone intensity scale compared to the standard gray. Such deviations from established convention can lead to ambiguity and unintended interpretations which makes it even more difficult for everyone to decipher precise meaning.

Heart rate data visualization is informative art that seeks to be and needs to be actionable. To do so, it is traditionally presented in three broad categories: numerical and textual (values), graphical and chart-based (histograms, radial graphs), and abstract and metaphorical (colors changes, shapes, gardens, flowers).

1. Numerical and Textual Displays: Usually represented as beats per minute or BPM, they are easy to read, easy to understand, but hard to contextualize, even for seasoned athletes or physiologists. Background information is needed such as age, maximum heart rate, level of training, current fitness level, etc. In short, they are less effective for deeper reflections

2. Graphical and Chart-based Visualizations. These are great for displaying both current and historical HR data. They show zones and levels of intensity through colors and bar height. They can also help estimate time. However, by themselves, they lack the granularity of numbers and require more time to look at. They are not a glanceable way to receive feedback on the go or during a hard session, especially when movement is involved.

3. Abstract and Metaphorical Visualizations. These approaches translate complex data into more relatable and evocative forms, usually to enhance engagement or reduce cognitive load. In doing so, they engage new users, children, and spark curiosity. However they lose the granularity of actionability. 

   
   

Regardless of the modality of delivery, physiological visualizations should be glanceable, simple, contextual, and actionable. The information should be presented to the user in a perceptible enough way to facilitate quick and almost reflexive responses. Most wearables or platforms currently presenting physiological data visually are either too crowded, too complicated, too abstract, too ambiguous, or too scientifically technical. They all lack real-time context. 

Current HR visualizations assume a physiological literacy that the average user may lack and place the burden of interpretation on said users. Not much is known about how these assumptions impact misinterpretation, inequitable access to insight, or decision-making during exercise, particularly among novices, diverse user groups, or in collaborative training contexts.

The main questions needing exploration

• What assumptions do current designs make about user literacy in exercise physiology?

• How much interpretation burden is pushed onto the user, rather than being supported by the system?

• Which visualization format or combination of formats (among the three listed above) reduce misinterpretation?

• To what extent do novice vs. experienced athletes interpret the same visualization differently?

• How does contextual information (weather, fatigue, illness, sleep) change interpretation of HR feedback? Heartbeat
  

My research fundamentally wants to establish interfaces which empower users to develop insights into their personal experiences. Such systems would make the distinction between contextualized situated information (integrating subjective and objective data such as goals, weather, or illness) and isolated raw physiological metrics (objective data alone). The core characteristic is to present information in a way that is glanceable and contextually meaningful (recovery or intervals – dry or raining – cold or hot). Finally, that data is to be presented in a way that makes processing effortless and suitable for the task or goal of the exercise session.

The ultimate goal is to remove any inference at the interface.