Open Epidemiology Platform Built for Africa
« La connaissance est comme un baobab : une seule personne ne peut l'embrasser. »
Wepisia is a complete architectural rewrite of OpenEpi, the world's reference open-source epidemiology calculator rebuilt with modern web technologies and extended with capabilities the original never had: multi-region epidemic simulation, interactive geospatial mapping, an in-browser R IDE, and multilingual support for African languages.
Everything runs locally in the browser. No server. No installation. No data leaves your device.
This project would not exist without the foundational work of the OpenEpi team. Their decision to release a complete, scientifically rigorous epidemiology toolkit under the MIT license made Wepisia possible. We owe them a profound debt.
OpenEpi Creators:
- Andrew G. Dean, MD, MPH
- Kevin M. Sullivan, PhD, MPH
- Roger Mir, MSc
- and all OpenEpi contributors
Primary Reference:
Dean AG, Sullivan KM, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health, Version 3.01. www.OpenEpi.com, updated 2013/04/06.
Peer-reviewed publication:
Sullivan KM, Dean AG, Mir R. OpenEpi: A Web-based Epidemiologic and Statistical Calculator for Public Health. Public Health Reports. 2009;124(3):471–474. DOI: 10.1177/003335490912400320
Every statistical result in Wepisia is validated against OpenEpi v3. Where the two tools differ on edge cases (small-sample CI approximations), the difference is documented and annotated in the interface.
OpenEpi (2002) Wepisia (2026)
───────────────── ─────────────────────────────────────────
Static HTML/JS → React 18 + TypeScript + Vite
Desktop-only → Mobile-first, fully responsive
English-only → 6 languages incl. Mooré, Wolof, Swahili
No simulation → Multi-region SIR/SEIR/SEIRD/SEIQRD engine
No mapping → Interactive geospatial + heatmap + clustering
No R runtime → Full R IDE via WebAssembly (WebR)
No literature → PubMed explorer with MeSH generator
No dark mode → System-adaptive dark/light theme
Biostatistics Suite with 19 modules, OpenEpi-validated
Full replication of OpenEpi's Screening Test module:
- Sensitivity, specificity, PPV, NPV, accuracy Wilson CIs
- LR+/LR− Katz log method
- Diagnostic odds ratio, Cohen's Kappa, Shannon entropy reduction, bias index
- Cutoff-specific and level-specific analyses
- ROC curve with AUC (Hanley–McNeil 1982 CI)
- 2×2 Fisher exact, mid-P, Mantel–Haenszel, Yates correction, OR, RR, risk difference, attributable fractions (Woolf CI)
- R×C Pearson χ², degrees of freedom, p-value (jStat exact), Cramér's V, expected frequencies
- One rate 5 CI methods: mid-P exact (Miettinen 1974), Fisher (Armitage 1971), Normal (Rosner), Byar (Rothman–Boice 1979), Rothman–Greenland
- Two rates Rate ratio, rate difference, Mantel–Haenszel χ², exact Poisson test
- Mantel linear trend test (χ² at 1 df)
- OR and RR per exposure level vs. reference, Woolf CIs
- Independent t-test: Student (equal variance) + Welch (Satterthwaite df) + Hartley F-test
- One-way ANOVA with Bartlett's test for variance homogeneity
- Mean CI (Student t), median/percentile CI (binomial method)
- Proportions (Kelsey, Fleiss ±continuity correction)
- Cohort/RCT, unmatched case-control, mean difference
- SMR (Standardized Mortality Ratio) 6 methods: Exact Poisson/χ², mid-P, Byar, Vandenbroucke, Rothman–Greenland, Pearson χ²
Epidemic Simulation Engine
Multi-region compartmental modeling with real-time interactive visualization:
Models: SIR · SEIR · SEIRD · SEIQRD
Regions: User-configurable network. Each region has its own population, compartments, and connections. Mobility parameter drives inter-regional transmission.
Interventions (cumulative, multiplicative effects): Lockdown (−60% β), Vaccination (−40% β), Distancing (−20% β), Masks (−15% β), School closures
Visualizations:
- 2D map (Leaflet) and 3D globe (Globe.gl) with proportional markers and mobility flows
- Epidemic curves: S, E, I, R, D compartments + Rt effective tracking
- Phase portrait (S–I space)
- Force-directed network graph with infection-rate-colored nodes
Pre-built scenarios (real disease parameters): COVID-19 · Ebola (West Africa 2014) · Seasonal flu · Measles · Meningococcal meningitis · Cholera
Comparison mode: Side-by-side simulation with overlay curves.
Geospatial Module
- CSV/Excel import with automatic lat/lng column detection
- Three visualization modes: individual markers, intelligent clustering, weighted heatmap
- Multi-layer dataset management with independent per-layer settings
- Timeline animation slider for temporal datasets
- GeoJSON export
- AI-powered epidemiological analysis: cluster summary, trend detection, risk scoring (building)
In-Browser R IDE (WebAssembly)
A full R environment running entirely in the browser, no server, no installation.
- Engine: WebR R compiled to WebAssembly
- Editor: CodeMirror 6 with R syntax highlighting, autocompletion, bracket matching, fold gutters, lint
- Multi-file tabs with context menu and auto-save
- Automatic plot capture, grid view, animation mode, PNG export
- Package management via
webr::install() - Interactive console with command history
PubMed Explorer
- Full-text search across 35M+ NCBI biomedical references
- MeSH generator: free-text → controlled vocabulary via PubMed's translation API
- Filters: date range, publication type, sort by relevance or date
- Favorites (up to 200) with RIS export (EndNode, Zotero, Mendeley)
| Language | Status | Notes |
|---|---|---|
| Français | Complete | Full UI + documentation |
| English | Complete | Full UI + documentation |
| Mooré | Experimental | Community review welcome |
| Wolof | Experimental | Community review welcome |
| Swahili | Experimental | Community review welcome |
| Haoussa | Experimental | Community review welcome |
Wepisia is the first epidemiology platform to offer multilingual support in indigenous African languages. All translations are open to community contribution and verification by native speakers.
client/
├── src/
│ ├── pages/
│ │ ├── biostatistics/ # 19 statistical modules
│ │ ├── simulation/ # Epidemic simulation dashboard
│ │ ├── geospatial/ # Interactive mapping
│ │ ├── explorer/ # PubMed search
│ │ └── Workspace.tsx # R IDE
│ ├── components/ # Shared UI components
│ ├── contexts/ # Theme, i18n
│ └── lib/ # Utilities, notifications
| Concern | Technology |
|---|---|
| Framework | React 18 + TypeScript |
| Build | Vite |
| Styling | Tailwind CSS v4 |
| Routing | Wouter |
| Charts | Chart.js + React-chartjs-2 |
| 3D Globe | Globe.gl |
| Maps | Leaflet + D3 |
| Statistics | jStat (exact distributions) |
| R runtime | WebR (WebAssembly) |
| Code editor | CodeMirror 6 |
| PDF export | jsPDF + jspdf-autotable |
| i18n | i18next |
| Animations | Framer Motion |
- No data collection. No analytics. No cookies.
- All computations are local, nothing leaves the browser.
- No login required. No account. No tracking.
- HTTPS delivery. MIT licensed source code.
git clone https://github.com/Xcept-Health/wepisia.git
cd wepisia/src/client
npm install
npm run devRequires Node.js 18+. No backend needed.
Contributions welcome, biostatisticians, developers, translators, public health practitioners.
[email protected]
[email protected]
github.com/Xcept-Health/wepisia
Priority areas: language verification (Mooré, Wolof, Swahili, Haoussa), additional modules, FETP-aligned workflows.
F. Ariel Shadrac Ouédraogo Medical student · Self-taught developer · Burkina Faso Founder, Xcept-Health open-source medical AI for Africa
MIT License, free to use, modify, and distribute, including commercially.
Copyright (c) 2026 F. Ariel Shadrac Ouédraogo / Xcept-Health
The statistical methodologies implemented in this software are derived from OpenEpi (© Andrew G. Dean, Kevin M. Sullivan, Roger Mir), also MIT licensed. Full license text in LICENSE.
Wepisia v0.1.0-beta · April 2026 · Made in Burkina Faso 🇧🇫
"Without OpenEpi, no Wepisia. Without Wepisia, fewer tools for Africa."