area, issue, severity, peopleAffected, resources, riskScore, action. Currently displaying 7 reports.Situation at a glance
Aggregated across all monitored districts in the last 24 hours.
Reports & risk trend
People affected
Recommended response order
Ranked by AI risk score combining severity, population, and available resources. Select a row for details.
| Area | Issue | Severity | People | Resources | AI Score | Action | Open |
|---|---|---|---|---|---|---|---|
| Rakiraki | Flash Flooding | Critical | 3,200 | 2 boats, 1 medical team | 94 | Deploy rescue teams immediately | |
| Ba | River Overflow | Critical | 2,450 | 3 boats, evacuation buses | 89 | Evacuate low-lying settlements | |
| Tavua | Road Closure / Landslide | High | 850 | 1 clearing crew | 76 | Dispatch clearing crew within 2 hrs | |
| Lautoka | Water Shortage | High | 5,100 | 4 water tankers | 71 | Distribute water to priority zones | |
| Nadi | Health Risk – Dengue Spike | Medium | 1,200 | Mobile clinic | 58 | Deploy vector control & clinic | |
| Labasa | Power Outage | Medium | 980 | Utility crew | 47 | Restore priority feeder lines | |
| Suva | Localized Flooding | Low | 320 | Municipal team | 32 | Monitor and drain hotspots |
Gemini AI Response Summary
Priority focus: Rakiraki and Ba are the highest-risk areas in the current 6-hour window. Rakiraki shows a 94/100 AI risk score driven by flash flooding along the Penang River and 3,200 residents in low-lying settlements with limited road access. Ba follows at 89/100 with river overflow already breaching two evacuation thresholds.
Recommended sequence: (1) Deploy rescue boats and a medical team to Rakiraki within 30 minutes, (2) begin coordinated evacuation of Ba's Nailaga and Yalalevu settlements, (3) pre-position water tankers for Lautoka before the shortage escalates.
Signals used: rainfall telemetry, community SMS reports, road-closure feeds, and hospital capacity. Confidence: High.
Google Cloud Data Layer
How raw community reports become ranked, AI-summarised decisions — powered by Cloud Storage, BigQuery, NVIDIA RAPIDS and Gemini.
Uploaded community incident CSV files land in a Cloud Storage bucket before ETL.
Emergency reports are loaded into a partitioned BigQuery table for fast SQL analytics.
GPU-accelerated priority risk scoring on millions of rows — 11.4× faster than pandas.
Generates plain-language response summaries and recommended next actions.
Ranked priorities, resources, and AI insight in a single decision-support view.
Real BigQuery objects backing the dashboard — table, rows, and a working query.
SELECT area, issue_type, severity, people_affected, resource_status, road_access, risk_score, recommended_action, created_at FROM `pacific_response_intelligence.incident_reports` ORDER BY risk_score DESC LIMIT 10;
Uploaded community incident CSVs land here before being loaded into BigQuery.
NVIDIA RAPIDS Acceleration
GPU-accelerated priority risk-scoring lets response teams process massive incident datasets in a fraction of the time — quicker triage, quicker decisions.
Priority risk-scoring and incident ranking on 2,000,000 emergency reports — NVIDIA RAPIDS/cuDF on a Tesla T4 GPU vs traditional pandas on CPU.
CPU dataframe processing on 2M reports.
GPU-accelerated processing on Tesla T4.
Data pipeline
From ground-truth community signals to actionable decisions.
Agent-Ready Decision Tools
Structured decision-support tools that can be used by AI agents, emergency systems, or future integrations.
These tools allow future AI agents or external response systems to access structured decision intelligence.
Returns ranked emergency reports sorted by AI risk score, with optional severity and limit filters.
Returns the current status, risk level, resource condition, and recommended action for a selected area.
Returns a Gemini-powered response summary and top-line emergency statistics.