Brain Exposome · 腦健康 Exposome 檢測
An educational, front-end-only brain-health self-check and interactive world map, grounded in the
2024 Lancet Commission on dementia's 14 modifiable risk factors plus cumulative air-pollution
(PM2.5) exposure. It estimates an educational "brain-age acceleration" reference value and visualises,
country by country, dementia prevalence, PM2.5, population aging, PAF (population attributable
fraction) of modifiable risk, and MCI / SCD (mild cognitive impairment & subjective cognitive
decline) prevalence.
All calculation happens in the browser — no personal data is uploaded or stored.
Deployed at https://brain-exposome.mattye.dev. Previously lived at
mattye.dev/projects/dementia-exposome/; that path 301-redirects here.
Tech stack
- Astro 6 (static output) · plain CSS · no JS framework
- Leaflet for admin-1 (county/province) choropleth maps
- d3-geo orthographic canvas globe for the national world layer
- Data pipeline in Python (
scripts/build_data.py)
Develop
npm install
npm run dev # http://localhost:4321
npm run build # → dist/
npm run preview
Data
Derived, ready-to-serve assets live in public/data/ (committed). Raw, non-redistributable inputs
(NCD-RisC / WHO GHO / GBD / WorldPop / publisher PDFs) live in scripts/_data_in/ (gitignored).
Regenerate the derived assets with:
python scripts/build_data.py
Sources, PAF math, and per-country provenance are documented in docs/dementia-exposome/ and cited
in-app (references §). Key backbone sources: Livingston et al. 2024 (Lancet Commission), NCD-RisC,
WHO Global Health Observatory, GBD 2021, World Bank, and Bai et al. 2022 (global MCI meta-analysis).
Layout
public/data/ derived, served data (dementia, exposome, pm25, aging, mci, scd, geo)
scripts/ build_data.py + helpers
src/pages/index.astro the single-page app
src/layouts/ src/components/ src/styles/
docs/dementia-exposome/ methodology, provenance, data-refresh + migration notes
All estimates are modelled and for education only — not clinical or diagnostic advice.
