Carlos D. Prado S.Personal analytics platform
MoodAtlas
A public, privacy-safe interface for exploring longitudinal mood, activity, wrist, EEG, and facial-signal analytics from curated data products.
- Role
- Product design, data architecture, frontend, and pipeline orchestration
- Period
- 2026
- Status
- Live public system
Challenge
What the work needed to resolve.
Combine sensitive, multimodal personal data into useful longitudinal views without exposing raw notes, private labels, or source media in a public build.
Approach
A system designed for scrutiny.
- 01
Defined typed JSON contracts between independent Python producers and the frontend.
- 02
Separated committed public artifacts from a token-gated, loopback-only private mode.
- 03
Built an atomic refresh contract with provenance, staleness checks, and privacy checks at source and bundle boundaries.
Inside
What the system actually produces.
The figures below are rebuilt in the portfolio from synthetic series, so they show the shape of each output without carrying any real data out of the project.
The calendar view keeps every day visible instead of collapsing the year into an average, so streaks, dips, and gaps stay legible.
- Steady41%
- Rising24%
- Volatile21%
- Dip14%
Days are classified into behavioural regimes and reported as shares, which is a description of the record — not a claim about what caused it.
Outcome