MarcosHealth
A personal analytics lab for wearable data: trends, context and AI-written readings of recovery and activity.
- Role
- Design & development
- Context
- Personal project
- Focus
- Data visualisation, trends, AI analysis
- Trends Long windows before single readings
- AI Descriptive summaries, never diagnoses
- Personal Built for my own training data
Interactive demo — Sample data — no medical interpretation
Fig. 03 Sample data — no medical interpretation
Overview
Wearables collect an enormous amount of data and explain very little of it. MarcosHealth is a personal application that gathers health and activity data from wearable devices and turns it into something that can actually be read.
It combines data visualisation, long-term trends and analysis by AI models to give context around recovery, activity and physiological metrics. It is meant as a training companion, not a medical tool.
The challenge
Daily numbers are noisy. A single bad night or an unusual reading means little on its own; what matters is the trend, and how different signals move together over weeks.
AI models are good at summarising, but health is a domain where overconfident language does harm. The analysis has to describe and contextualise, never diagnose.
Architecture
Key decisions
- 01
Trends over snapshots
Visualisations favour rolling averages and long windows over single days, so the shape of the data is visible before any individual point.
- 02
Context, not conclusions
AI-written summaries describe what changed and how it relates to other metrics. They avoid diagnoses, medical advice and any certainty the data cannot support.
- 03
A calm, premium interface
The interface borrows from sports analytics: generous spacing, restrained colour and charts designed to be understood in seconds.
Stack
- Wearable data
- Data visualisation
- Trend analysis
- AI models
Notes
All figures on this site are sample data generated for the portfolio. MarcosHealth does not provide medical diagnoses or advice.