SlopeSense Web | GitHub Repository
The predictive Landslide Intelligence Platform built to save lives. SlopeSense is a pure-software early warning system that fuses free satellite data into a probabilistic Failure Probability Index (FPI). Computed at 1km² resolution and updated every 6 hours, it delivers critical warnings directly to decision-makers with a 24–48 hour lead time before disasters strike.
Experience how SlopeSense delivers actionable insights through an intuitive GIS dashboard. The platform aggregates complex geospatial data into a simple, interactive format.
India loses ~800 lives per year to landslides. On July 30, 2024, Wayanad saw 420 deaths — a warning existed 16 hours prior but was never integrated into the official channel.
The gap is not science. It is the operational intelligence layer between raw satellite data and the district collector who needs to order an evacuation.
SlopeSense fuses free satellite data from 6 sources into a probabilistic Failure Probability Index (FPI) — computed at 1km² resolution, updated every 6 hours — and delivers it directly to decision-makers.
| Capability | Detail |
|---|---|
| Risk Model | Physics-based FPI (derived from NASA LHASA v2) + LightGBM calibration |
| Resolution | 1km² grid cells → block-level aggregation |
| Update Frequency | Every 6 hours during monsoon periods |
| Forecasting | 24–48 hour forward forecast using NOAA GFS |
| Alert Delivery | GIS dashboard (MapLibre GL) + WhatsApp Business API |
| Standards | CAP v1.2 XML feed (NDMA Sachet-compatible) |
| Audit Trail | Full retrospective validation for every alert issued |
+-----------------------------------------------------------------------------+
| SlopeSense Platform |
| |
| +------------+ +--------------+ +-----------+ +-------------------+ |
| | Satellite | | Preprocessing| | FPI Model | | Alert Dispatch | |
| | Ingestion |->| + Regridding |->| Engine |->| + CAP Feed | |
| |(GPM, SMAP, | |(numpy+xarray)| |(LHASA v2) | | + WhatsApp | |
| | Sentinel-2) | +--------------+ +-----------+ +-------------------+ |
| +------------+ | |
| v |
| +-------------------------------------------------------------------------+|
| | FastAPI REST + WebSocket API ||
| | /v1/alerts /v1/risk /v1/geojson/fpi /v1/cap/feed /ws/live ||
| +-------------------------------------------------------------------------+|
| | |
| +-------------------------------------------------------------------------+|
| | Next.js 14 Dashboard (MapLibre GL JS) ||
| | FPI Heatmap Block Risk Table Alert Feed CAP Viewer ||
| +-------------------------------------------------------------------------+|
+-----------------------------------------------------------------------------+
See ARCHITECTURE.md for full system diagrams with Mermaid.
slopesense/
├── backend/ # Python / FastAPI backend
│ ├── api/ # REST API application
│ │ ├── main.py # All route handlers
│ │ ├── database.py # SQLAlchemy async engine
│ │ ├── middleware.py # Auth, rate limiting, logging
│ │ ├── apikeys.py # API key management routes
│ │ ├── cache.py # In-memory TTL cache
│ │ ├── metrics.py # Prometheus metrics
│ │ ├── reports.py # PDF report generation
│ │ └── webhooks.py # WhatsApp webhook handler
│ ├── ingestion/ # Satellite data fetchers
│ │ ├── gpm.py # NASA GPM IMERG (rainfall)
│ │ ├── smap.py # NASA SMAP L3 (soil moisture)
│ │ ├── copernicus.py # ESA Copernicus DEM + Sentinel-2
│ │ └── open_meteo.py # NOAA GFS / Open-Meteo forecasts
│ ├── processing/ # Data preprocessing pipeline
│ │ └── preprocessor.py # Regridding, slope, percentiles
│ ├── model/ # FPI inference engine
│ │ ├── fpi_engine.py # LHASA v2 implementation + labels
│ │ ├── retrospective.py # Historic event validation
│ │ └── train.py # LightGBM calibration training
│ ├── alert/ # Alert generation and dispatch
│ │ ├── alert_engine.py # Threshold logic, CAP XML, WhatsApp
│ │ ├── dispatcher.py # Async delivery (WhatsApp + email)
│ │ └── verifier.py # HMAC webhook verification
│ ├── migrations/ # Alembic database migrations
│ ├── tests/ # Pytest test suite (100+ tests)
│ │ ├── conftest.py # Fixtures and in-memory DB setup
│ │ ├── test_api.py # API endpoint integration tests
│ │ ├── test_fpi.py # FPI engine unit tests
│ │ ├── test_cache.py # Cache layer tests
│ │ ├── test_middleware.py # Auth + rate limiting tests
│ │ └── test_risk_labels.py # Risk label semantic tests
│ ├── models.py # SQLAlchemy ORM models
│ ├── config.py # Pydantic settings (env-driven)
│ ├── worker.py # Celery task definitions
│ ├── requirements.txt # Python dependencies (pinned)
│ ├── Dockerfile # Backend container image
│ └── alembic.ini # Alembic migration config
│
├── frontend/ # Next.js 14 dashboard
│ ├── src/
│ │ ├── app/ # App Router pages
│ │ ├── components/ # React components
│ │ │ ├── Navigation.tsx # Persistent nav header
│ │ │ ├── dashboard/ # Dashboard panel components
│ │ │ └── map/ # MapLibre GL map components
│ │ └── lib/
│ │ ├── api.ts # Typed API client
│ │ └── hooks.ts # React data-fetching hooks
│ ├── Dockerfile # Frontend container image
│ └── package.json
│
├── infra/ # Infrastructure configuration
│ └── nginx.conf # Nginx reverse proxy rules
│
├── docs/ # Extended documentation
│ ├── API.md # Full API reference with examples
│ ├── DATA_SOURCES.md # Satellite data source details
│ ├── BUSINESS.md # Business model and pricing
│ ├── DEPLOYMENT.md # Production deployment guide
│ └── sample_cap.xml # Sample CAP v1.2 XML output
│
├── scripts/ # Utility and maintenance scripts
├── data/ # Local data cache (gitignored)
├── docker-compose.yml # Full stack orchestration
├── docker-compose.override.yml # Local development overrides
├── pytest.ini # Pytest configuration
├── .env.example # Environment variable template
├── README.md # This file
├── ARCHITECTURE.md # System architecture + diagrams
├── CHANGELOG.md # Version history (Keep a Changelog)
└── CONTRIBUTING.md # Contributor guide
| Tool | Version | Purpose |
|---|---|---|
| Docker + Docker Compose | 24+ | Container orchestration |
| Python | 3.11+ | Backend development |
| Node.js | 18+ | Frontend development |
| NASA Earthdata account | — | GPM + SMAP data (free) |
| ESA Copernicus account | — | DEM + Sentinel-2 (free) |
# 1. Clone and configure
git clone https://github.com/slopesense/slopesense.git
cd slopesense
cp .env.example .env
# Edit .env with your API credentials
# 2. Start the full stack (DB -> API -> Worker -> Frontend -> Nginx)
docker-compose up -d
# Dashboard: http://localhost
# API: http://localhost/api
# API Docs: http://localhost/api/docs# Backend
cd backend
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
# Start infrastructure (PostgreSQL + Redis)
docker-compose up -d db redis
# Run database migrations
alembic upgrade head
# Start API server with hot-reload
uvicorn backend.api.main:app --reload --host 0.0.0.0 --port 8000
# Frontend (new terminal)
cd frontend
npm install
npm run dev # http://localhost:3000# Backend — unit + integration tests (100+ cases)
pytest backend/tests/ -v --cov=backend --cov-report=html
# Frontend — lint + TypeScript type checking
cd frontend
npm run lint
npx tsc --noEmit| Source | Data Type | Spatial Res. | Update Latency |
|---|---|---|---|
| NASA GPM IMERG | Rainfall | 0.1 deg (11km) | 4 hours |
| NASA SMAP L3 | Soil Moisture | 36km | 24-48 hours |
| Copernicus DEM GLO-30 | Elevation/Slope | 30m | Static |
| ESA Sentinel-2 L2A | NDVI (vegetation) | 10m | 5 days |
| NOAA GFS | Forecast Rainfall | 0.25 deg (28km) | 6 hours |
| NDMA NLSM | Susceptibility Prior | District | Annual |
| Tier | FPI Range | Recommended Action |
|---|---|---|
| Normal | less than 40% | Routine monitoring |
| Watch | 40-65% | Alert DDMA. Heighten awareness. |
| Warning | 65-80% | Pre-position NDRF/SDRF. Issue public advisory. |
| Emergency | greater than 80% | Immediate evacuation advisory. |
| Monitoring | Any (suppressed) | High model uncertainty — observe only. |
Full documentation at /docs (Swagger UI) and /redoc. See also docs/API.md.
| Method | Endpoint | Auth | Description |
|---|---|---|---|
| GET | / |
None | Health check + system status |
| GET | /v1/risk |
None | FPI for a lat/lon point |
| GET | /v1/alerts/active |
None | Current active alerts |
| GET | /v1/alerts/{id} |
None | Alert detail with signal breakdown |
| GET | /v1/districts/{state} |
None | All districts with current FPI |
| GET | /v1/blocks/{district} |
None | Block-level FPI scores |
| GET | /v1/historical/{date}/{district} |
None | Historical FPI |
| GET | /v1/retrospective |
None | Validation audit results |
| GET | /v1/cap/feed |
None | CAP v1.2 XML feed (Sachet-compatible) |
| GET | /v1/geojson/fpi |
None | GeoJSON FPI grid for MapLibre |
| POST | /v1/contacts/register |
API Key | Register for WhatsApp alerts |
| WS | /ws/live |
None | Real-time WebSocket feed |
| Metric | Value |
|---|---|
| API P95 response latency | less than 120ms |
| Full model run (all-India) | ~4 minutes |
| Alert-to-WhatsApp delivery | less than 30 seconds |
| Retrospective accuracy (T-24h) | 6/6 historic events flagged |
| Revenue Stream | Target Value |
|---|---|
| SDMA SaaS state contracts | Rs. 15-25 L/year per state |
| NDMA national contract | Rs. 1-3 Cr/year |
| World Bank / UNDP DRR grants | Rs. 2-5 Cr (one-time) |
| Open Data API paid tier | Research, NGOs, reinsurers |
See docs/BUSINESS.md for the full commercial model.
We welcome contributions from developers, data scientists, and disaster management practitioners. Please read CONTRIBUTING.md before submitting a PR.
Apache 2.0 — same as NASA LHASA v2, the base model this project builds on.
"On July 29, 2024, SlopeSense had a 73% risk score for Meppadi block with a forward forecast of 81% for the following 24 hours. The actual landslide occurred at 2:17am on July 30. The Gram Pradhan would have received this WhatsApp message 20 hours earlier."