How siddhant0812-sid/aquawatch-flood works
AquaWatch Flood is a local flood-monitoring application for the Assam/Brahmaputra basin. Its main user workflow loads station status and SAR flood mapping, selects a station for 24/48/72-hour forecasts, and can simulate an emergency alert; separate views expose station details and model evaluation. Internally, hydrological telemetry is cleaned and transformed into forecast features, while SAR scenes are segmented and vectorized for map display. The graph preserves the documented ARIMA/LSTM/TFT comparison and explainability capabilities. Some backend and evaluation links rely on README documentation because their implementation excerpts were not provided.
Main entry points: DashboardPage.tsx Dashboard View · App.tsx App Router · model_lstm.py Multi-horizon LSTM · explainability.py Comparison & Explainability
Components 21
Dashboard UI
- App Router
frontend/src/App.tsx - Dashboard View
frontend/src/pages/DashboardPage.tsx - Station Detail
frontend/src/pages/StationDetailPage.tsx - Evaluation View
frontend/src/pages/ModelEvaluationPage.tsx - Basin Map
frontend/src/components/dashboard/BasinMapPanel.tsx - KPI Metrics
frontend/src/components/dashboard/KpiMetricsRow.tsx - SAR Showcase
frontend/src/components/dashboard/SarShowcasePanel.tsx - Forecast Chart
frontend/src/components/dashboard/ForecastChartPanel.tsx - Hydrological Drivers
frontend/src/components/dashboard/HydrologicalDriversPanel.tsx - Station Table
frontend/src/components/dashboard/StationTablePanel.tsx
Application API
- Typed API Client
frontend/src/api/client.ts - FastAPI Endpoints
backend/main.py - Model Service
backend/model_service.py
SAR Flood Mapping
- U-Net Inference
models/inference.py
Hydrological Forecasting
- Telemetry Pipeline
models/data_pipeline.py - Feature Engineering
models/feature_engineering.py - Multi-horizon LSTM
models/model_lstm.py
Evaluation & Explainability
- ARIMA Baseline
models/baseline_arima.py - TFT-Lite Model
models/model_tft.py - Comparison & Explainability
models/explainability.py
Other components
- Flood Operator
Connections 24
- Flood Operator → App Router: uses
README.md - App Router → Dashboard View: routes to
frontend/src/App.tsx - App Router → Station Detail: routes to
frontend/src/App.tsx - App Router → Evaluation View: routes to
frontend/src/App.tsx - Dashboard View → Typed API Client: requests data
frontend/src/pages/DashboardPage.tsx - Dashboard View → Basin Map: renders
frontend/src/pages/DashboardPage.tsx - Dashboard View → KPI Metrics: renders
frontend/src/pages/DashboardPage.tsx - Dashboard View → SAR Showcase: renders
frontend/src/pages/DashboardPage.tsx - Dashboard View → Forecast Chart: renders
frontend/src/pages/DashboardPage.tsx - Dashboard View → Hydrological Drivers: renders
frontend/src/pages/DashboardPage.tsx - Dashboard View → Station Table: renders
frontend/src/pages/DashboardPage.tsx - Typed API Client → FastAPI Endpoints: calls endpoints
README.md - FastAPI Endpoints → Model Service: delegates to
backend/main.py - Model Service → U-Net Inference: runs segmentation
backend/model_service.py - Model Service → Multi-horizon LSTM: runs forecasting
backend/model_service.py - Telemetry Pipeline → Feature Engineering: provides telemetry
models/feature_engineering.py - Feature Engineering → Multi-horizon LSTM: builds model inputs
models/model_lstm.py - ARIMA Baseline → Telemetry Pipeline: uses pipeline
models/baseline_arima.py - ARIMA Baseline → Feature Engineering: uses features
models/baseline_arima.py - TFT-Lite Model → Multi-horizon LSTM: reuses model code
models/model_tft.py - Comparison & Explainability → ARIMA Baseline: compares
README.md - Comparison & Explainability → Multi-horizon LSTM: explains
README.md - Comparison & Explainability → TFT-Lite Model: compares
README.md - Evaluation View → Typed API Client: loads evaluation
README.md