Assam Basin Flood Outlook
Nine river reaches · Brahmaputra & Barak basins · 24–48 h probabilistic lead
Catchment risk map
Assam district boundaries · 9 gauging reachesRisk ranking
calibrated probability × 100Alert dispatch log
hindcast validation · last 14 daysMonsoon season timeline
mean risk across 9 reaches · bars = actual exceedances · click to jumpMonitored catchments
select a card to open station intelligence →Poore Assam ka direct-warning grid
Sivasagar (19 July 2026) ko calibration event banaya gaya hai, par drill kisi bhi upstream zone — Naga Hills, Arunachal, Bhutan foothills, Meghalaya ya Barak headwaters — se chalaya ja sakta hai. Har downstream district ko ek CAP alert aur multi-channel direct-warning plan milta hai.
Drill console
surge kahan se aayegi? · kaunsa sensor tutega?Trigger & fan-out timeline
15-min reconstructed telemetry · T+ minutes from 06:00 ISTCAP 1.2 message (direct warning payload)
Test status · operator ACK required · NOT DISPATCHEDDirect-warning channels
send ≠ delivered ≠ acknowledgedDelivery / ACK ledger
1,000 synthetic recipients per district · no real numbersHardware-ready ingest API
ESP32 / LoRa / GSM node yahi JSON post karegaAssam-wide fan-out map
33 districts · 5 upstream sentinels · colour = drill severityWhy a state-wide grid?
Sivasagar ka trigger Assam ke andar nahi tha19 Jul 2026 · Sivasagar replay
reconstructed hydrograph, NOT official 15-minute CWC telemetrySensor → siren architecture
edge autonomy, fail-safe degradation, operator acknowledgementHardware & packet contract
design references, not a deployed bill of materialsSources & honest caveats
real facts vs synthetic drillDoes machine learning actually help?
Every number below is computed on 2025–2026 monsoon days the model never saw during training, threshold tuning or calibration.
Benchmark against simple rules
same test set, same label, same threshold policyROC curve
Precision–Recall
Calibration
Confusion matrix
Accuracy vs latency — the real trade-off
why a great AUC is useless if the input data is 2–5 days oldWhat drives a prediction?
permutation importance: drop in test ROC-AUC when one feature is shuffledReal event window
rainfall, risk score and observed exceedancesRisk by station × month
System architecture
from public data to a served predictionModel card & honest limitations
what this system can and cannot doFrom a static demo to a
reproducible ML system
Reproducibility, provenance & limits
Everything needed to rerun the pipeline, inspect the data and explain the project in an interview.