OVERVIEW
Out-of-the-box health visualizations, say from a health tracking app, often leave out life contexts which can be important to surface when reviewing data. For instance, a rapid increase in resting heart rate can could be cause for alarm — or not if the individual is pregnant. Through interviews and visual elicitation, I explored how women post-pregnancy reflected on data from their pregnancy to understand how these data should be represented back to them now that pregnancy is over.
THE PROBLEM
Disruptions to routine life are inevitable. Take, for example, a pandemic. During a pandemic, if you wore a wearable sensor such as an Apple Watch, it would show that you had taken very few steps during the year 2020. When looking back at your data in later years, knowing there was a pandemic going on would help you make sense of this low step count. Yet, information on disruptions are often absent from data visualizations. Data visualizations that do not reflect contexts of disruption may lead to data misinterpretation and missed opportunities to gain an understanding about one’s own health. Pregnancy can be considered another type of disruption. In a two-part study, I seek to answer the questions:
(1) HOW DO self-trackers use
pregnancy data after pregnancy?
and
(2) How Could health data visualizations
reflect life disruptions?
PROCESS
PARTICIPANTS
All the participants were women who had tracked stress during their pregnancy by wearing a heart rate monitor and by answering ecological momentary assessments about their stress (EMAs) using their phone.
PROCEDURE
1-hour individual remote interviews with 8 mothers using visual elicitation methods
Participants were asked about their prior experiences tracking their own health
Visual elicitation exercise (see screenshot below) in order to ground the remainder of the interview through a mutual understanding of the context in which the pregnancy was situated
Textual elicitation exercise (see screenshot below) Participants were probed on whether they would still find different types of data captured during their pregnancy relevant and why
Qualitative analysis
PRELIMINARY FINDINGS
Data that was captured during a disruptive period (e.g. pregnancy) is indeed integral to data interpretation once the context has ended and routine has returned.
Data from pregnancy was used to define the boundaries of pregnancy. Boundaries of life contexts such as pregnancy are subjective and can be define by the data (e.g. Pregnancy is not over until my sleep is back to ‘normal’)
Individuals review data from a past pregnancy when the data or insights are still applicable to non-pregnancy routine life. Data on weight may no longer be relevant post-pregnancy, but data on stress may still be relevant. How can we design visualizations that take into account how different types of data vary in usefulness over time?
Data from a prior pregnancy can be used to model a future pregnancy. Individuals can learn and self-experiment by viewing data from a prior-self. How can we design visualizations that support learning about the self during disruptive periods?
NEXT STEPS — !* This is an ongoing study *!
What if individuals were able to highlight or hide streams of data they believe were impacted by a disruption? How do individuals decide which data they choose to keep and which they find irrelevant? What are the benefits and challenges to manipulating their health data visualizations?
In the next phase of this research, I investigate how mothers might want to manipulate their health data visualizations from during their pregnancy.
PRESENTATION
This work was accepted to the 2020 Pervasive Health Doctoral Consortium. Doctoral consortia are workshops typically held at conferences intended for a select group of PhD students to receive feedback on their work. See my submission titled ‘Designing Visualizations of Self-Tracked Data From Bounded Situational Contexts: An Exploratory Study of Stress Data Captured During Pregnancy.’