Overview of The Lived Informatics Model
In this episode of Data Skeptic, host Kyle Polich speaks with Daniel Epstein, associate professor of Informatics at UC Irvine, about human-computer interaction and the evolution of self-tracking research into what Epstein calls the lived informatics model. The conversation explores how wearable devices and tracking apps are used in everyday life, why people adopt them, why they stop using them, and how the field is adapting to more complex, realistic human goals—especially as AI, health data, and privacy concerns reshape the landscape.
What “Informatics” and “Personal Informatics” Mean
Informatics
- Broadly covers how people interact with technology in everyday life.
- Focuses on designing, understanding, and improving technology’s role in society.
- Draws on methods like:
- interviews
- prototyping
- usability studies
- fieldwork
- system building and evaluation
Personal Informatics
- Essentially self-tracking: collecting data about yourself to better understand behavior, health, or routines.
- Common examples include:
- Apple Watch
- Fitbit
- Whoop
- sleep and step tracking
- food journaling apps
- The field looks at both:
- measurement and dashboards
- reflection and insight that help people act on the data
The Lived Informatics Model
Epstein explains that the lived informatics model emerged after the rise of mainstream wearables and mobile tracking apps changed self-tracking from a niche “Quantified Self” activity into something ordinary people do.
Why it mattered
- Earlier self-tracking was often associated with highly engaged “biohackers” who built custom tools and analyzed their own data.
- The new wave of wearables made tracking accessible to everyone, so researchers needed a model that reflected real-world, everyday use rather than idealized, highly technical users.
Core idea
- The model reframes tracking as something embedded in ordinary life, where goals, motivation, burden, and context all matter.
- It is less about perfect optimization and more about how people actually use tracking over time.
Why People Track
Epstein describes three common motivations:
- Behavior change
- Example: walking more, exercising more, eating healthier
- Instrumental tracking
- Example: record-keeping, insurance requirements, or other practical needs
- Curiosity
- People try a device or app without a clear long-term goal, often because they got it as a gift, recommendation, or promotion
What the Research Shows About Use and Drop-Off
Engagement is often temporary
- Many people stop using trackers after weeks or months.
- This is not always a failure—often it means they got what they needed.
Different motivations lead to different usage patterns
- Behavior-change users tend to track more consistently and for longer.
- Curiosity-driven users often experiment briefly, switch tools, or abandon tracking quickly.
Tracking can also have downsides
- Data can create judgment, guilt, or fixation.
- Food tracking, in particular, can make people feel like they are failing if the numbers don’t match their goals.
- In some cases, tracking becomes overly obsessive and may harm mental or physical well-being.
What Success Looks Like in This Field
From a research perspective, success is not “retention at all costs.”
Better definition of success
- People gain self-understanding
- They reflect on their habits
- They use the data to make decisions
- They eventually stop tracking when it is no longer useful
This creates tension with product/business goals, where companies often want long-term engagement and subscriptions.
Good Ecosystems for Tracking Data
Epstein argues that a healthy ecosystem should let users:
- export their data
- move it across devices
- preserve long-term history
- avoid being locked into one brand or product
Why continuity matters
Long-term data can reveal meaningful patterns, such as:
- changes after having a child
- shifts after moving from suburb to city
- long-term trends in activity or health
AI’s Role in Self-Tracking
AI appears to be especially promising in two areas:
1. Finding insights in messy data
- AI can help analyze large, heterogeneous datasets.
- It may surface trends that are hard for users to spot on their own.
2. Building personalized tracking tools
- AI could help people create apps tailored to what matters to them.
- Example: a chatbot-style tool that helps define what to track and at what level of detail.
Caution
- Epstein is optimistic, but also wary of putting highly personal data into AI systems.
- Privacy, data use, and downstream implications remain major concerns.
Research Examples Mentioned
Epstein describes several domains his group studies:
- Baby tracking
- feeding, sleep, diapers, memory support for sleep-deprived parents
- Period and fertility tracking
- Food journaling
- especially when goals shift, such as from weight loss to weight maintenance or higher protein intake
- Sustainability / carbon tracking
- including the footprint of LLM usage
- GLP-1-related food tracking
- how people using drugs like Ozempic interact with food logs
- Medical-grade personal informatics
- such as continuous glucose monitors or sleep devices
Baby Tracking as a Special Case
Baby tracking challenged some assumptions in the lived informatics model.
Key difference
- It is often less about behavior change and more about memory augmentation.
- Parents are tracking to remember:
- when the baby last ate
- when the baby last slept
- when a diaper change happened
Implication
- The time scale is much shorter than in other forms of self-tracking.
- This showed that not all tracking is aimed at long-term reflection; some of it is about immediate survival and coordination.
Medical Use and Data Sharing Challenges
Epstein notes that medical-grade tracking is also part of personal informatics, but it introduces extra complexity.
Challenges
- Doctors have limited time to review patient data
- Patients may have far more detailed logs than clinicians can easily use
- Translating personal data into something useful in a medical setting remains an open problem
Privacy and Data Governance
Epstein says there are valid reasons to be cautious about personal tracking data.
Main concerns
- health data can affect insurance and other systems
- companies may not give users full access or control
- data may not be portable across platforms
His view
- The risks are real, but many are similar to privacy risks in other digital spaces
- Health data can be especially sensitive, so skepticism is warranted
Key Takeaways
- Self-tracking is not just about optimization; it is about how technology fits into lived experience.
- People track for different reasons: behavior change, practical record-keeping, or curiosity.
- Dropping a tracker is often a sign of success, not failure.
- A strong ecosystem should support portability and long-term continuity of data.
- AI could improve insight generation and app creation, but privacy and misuse are important concerns.
- New research is shifting from single-goal tracking to the messier reality of overlapping life priorities.
Further Reading / Follow-Up
Daniel Epstein
- Website:
d-epstein.net - Also active on:
- Google Scholar
- Bluesky / X intermittently
Episode theme
- A strong introduction to how HCI and informatics research are thinking about self-tracking in the age of wearables, apps, and AI.