SIGMA

Spatial Intelligence & Geographic Modeling for Action

SIGMA is a public-health intelligence and scenario-analysis platform concept for a hackathon project. It combines formal health data and open-source public-health signals, identifies spatiotemporal patterns and anomalies, and provides interactive analysis and scenario-based decision support.

Status: early build. A dengue data pipeline (cleaning, DOH alert and epidemic thresholds, EARS C1/C2/C3 anomaly detection) is implemented in ml/; see docs/dengue-pipeline.md. A versioned JSON API (/api/v1) and an Expo client for it are in place; see docs/api.md and docs/mobile.md. The anomaly engine is implemented but not connected to the API, the ACCESS module runs on simulated records, and the Policy Copilot needs a server-side model provider key. Nothing here is a validated public-health claim.

Core modules

Module	Route	Purpose
TRACE	/lab	Detect and organize public-health signals.
ANALYZE	/explore	Explore trends, anomalies, spatial clusters, historical data.
SIMULATE	/simulate	Explore simple spatial-temporal scenarios.
ACT	/	Present evidence and decision-support info for LGU/public-health users.

Project structure

D:\SIGMA\
  web\                Next.js + TypeScript + Tailwind frontend (App Router)
  mobile\             Expo + React Native client (iOS, Android)
  packages\
    sigma-contracts\  Type-only API contracts shared by web and mobile
  ml\                 Python environment and ML/analysis code
    .venv\            Local virtual environment (not committed)
  data\
    raw\              Raw, unprocessed data (not committed)
    processed\        Processed/derived data (not committed)
    sample\           Small sample data safe to commit for development
  scripts\            Automation / utility scripts
  docs\               Project documentation
  .github\workflows\  CI validation workflows
  .cache\             Local caches (not committed)

Tech stack

• Frontend: Next.js, TypeScript, Tailwind CSS, App Router
• Mobile: Expo SDK 58, React Native, expo-router, EAS builds
• Deployment: Vercel (not yet configured)
• Database: Supabase PostgreSQL (not yet configured)
• Maps: MapLibre GL JS, OpenStreetMap-compatible data
• Charts: Recharts (web), plain React Native views (mobile)
• Python / ML: numpy, pandas (plus openpyxl, matplotlib and reportlab for the dengue pipeline and its QA reports)
• Automation: GitHub Actions

Getting started

Frontend

cd web
npm install
npm run dev

Then open http://localhost:3000.

Routes: /, /explore, /lab, /simulate, /policy, /access.

Mobile

cd mobile
npm install
npm start

The app talks to the web API. The address is set in Settings inside the app, or baked in with EXPO_PUBLIC_API_BASE_URL. The app uses the SIGMA field-guide design system — warm paper, ink, cobalt, olive, terracotta — defined in mobile/src/constants/theme.ts and documented in docs/mobile.md, together with the EAS build profiles and the APK install steps. Sigi, the maya-inspired companion, is generated pixel-art (web/scripts/generate-sigi-mascot.py) and appears on the splash and in the Overview, Dengue, ACCESS and Policy screens; it states what the app is doing, never what the data means. The package identifier is still a placeholder (org.sigma.mobile), and the API is not deployed, so an APK can only reach a server on the same network.

API

The web app serves a versioned JSON API under /api/v1. See docs/api.md.

Endpoint	Method	Returns
/api/v1	GET	The endpoint index
/api/v1/health	GET	Service status and configuration flags
/api/v1/sources	GET	The source catalogue
/api/v1/dengue/observations	GET	Raw weekly national observations
/api/v1/dengue/summary	GET	Weekly, monthly and yearly aggregates
/api/v1/dengue/signals	GET	Anomaly results and open-source demo data
/api/v1/access/metrics	GET	Medication access indicators
/api/v1/policy/info	GET	Assessment areas and the AI disclosure
/api/v1/chat	POST	The Policy Copilot

Python / ML environment

cd ml
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt

Environment variables

Copy the example files to .env.local (frontend), .env (ml) and .env (mobile) and fill in the values you need. No real credentials are included in this repository.

• web/.env.example — the model provider key (server-side only) and the public map-tile settings.
• ml/.env.example — database connection details, still empty because there is no database yet.
• mobile/.env.example — the API base URL. This is the only thing that may be set in the app, and it must never be a secret.

Checks

cd web    && npm run lint && npm run typecheck && npm run test && npm run build
cd mobile && npm run lint && npm run typecheck && npm run test
python -m compileall -q ml

Documentation

Document	Covers
docs/architecture.md	The high-level system architecture
docs/api.md	The /api/v1 endpoints, envelopes, limits and error codes
docs/mobile.md	The Expo client, the API address, and the EAS build and install steps
docs/dengue-pipeline.md	The dengue cleaning and EARS pipeline (QA reports in docs/reports/)
docs/access-data-model.md	The ACCESS data model, which is design-only so far
docs/data-sources.md	What each data source is and what it is not

Notes

• The vintage/illustrated watercolor SIGMA logo is a brand asset only (web/public/brand/) and is not used as the application's UI style.
• The application UI itself is modern, clean, scientific, and dashboard-oriented.
• This is a hackathon scaffold. Nothing in this repository should be treated as a validated public-health claim or production-ready system.