InitAlert PH: Hyperlocal Heat Index AI for School Safety & Outdoor Labor Protection

Problem Statement

During El Niño dry seasons, Philippine heat indices frequently breach the 42°C to 48°C “Danger” threshold across Luzon, Visayas, and Mindanao. Public school classrooms—frequently crowded with 45–50 students under uninsulated galvanized iron roofs with minimal electric fans—turn into hazardous thermal ovens, leading to widespread student fainting, epistaxis (nosebleeds), heat exhaustion, and hospitalizations. Mayors, school principals, and DepEd Division Superintendents face chaotic, last-minute cancellation decisions, often announcing class suspensions when students are already halfway to school or under scorching midday sun. Similarly, hundreds of thousands of outdoor construction and agricultural workers face deadly heat stroke without mandatory rest-and-hydrate protocols.

Proposed Solution

InitAlert PH is an AI-powered thermal risk intelligence system designed to protect vulnerable communities from extreme El Niño heat waves.

1. Hyperlocal Predictive Heat Modeling: Ingests PAGASA synoptic station forecasts, Open-Meteo microclimate feeds, and Sentinel-3 Land Surface Temperature (LST) data. It combines ambient temperature and relative humidity with building characteristics (corrugated iron roofs, lack of urban tree canopy, floor level) to model the internal thermal index of individual school campuses and construction corridors 48 hours ahead.
2. Automated DepEd / LGU Policy Generator: When the predicted heat index exceeds 42°C, the AI drafts ready-to-sign executive advisories for City Mayors and DepEd Superintendents, recommending pre-emptive transitions to asynchronous/distance learning (ADM) with clear parent notifications sent the evening before.
3. Smart Worker Heat Shield: Delivers automated SMS advisories to site engineers and outdoor workers with dynamic “work-rest ratios” (e.g., 30 minutes work, 30 minutes shade/hydration per hour) in compliance with DOLE Labor Advisory No. 08-2023.
4. Public Heat Risk Map: A mobile-friendly dashboard for parents, jeepney commuters, and senior citizens showing real-time street-level heat safety ratings and nearby public cooling centers.

Target Market

• Primary Beneficiaries: 28+ million Filipino public school students, teachers, outdoor laborers, delivery riders, and elderly citizens vulnerable to heat stroke.
• Primary Buyers (B2G / B2B): Department of Education (DepEd), Local Government Units (1,600+ City and Municipal Mayors’ Offices), Department of Labor and Employment (DOLE), Occupational Safety and Health Center (OSHC), and large construction/logistics conglomerates.

How AI Is Used

• Microclimate & Urban Heat Island (UHI) Machine Learning: Spatio-temporal regression models correlate macro weather stations with satellite thermal imagery and elevation maps, downscaling broad municipal heat alerts to 100-meter neighborhood grids.
• Indoor Heat Index Simulation: Neural networks approximate indoor radiant temperature based on outdoor meteorological parameters, building envelope materials, and classroom occupancy density.
• Executive Policy Drafting: LLMs auto-generate official administrative orders, suspension announcements, and parent SMS broadcasts in formal Filipino and English, adhering to DepEd Order No. 037 guidelines.
• Why it fails without AI: Broad regional forecasts from a single PAGASA synoptic station (e.g., Science Garden in QC) cannot reflect microclimate variations between tree-covered university campuses and concrete, tin-roofed community high schools across a sprawling metropolis.

PLDT / Smart / ePLDT / Rokid Angle

• Smart Communications (Emergency Cell Broadcast & SMS): Smart’s network broadcasts localized heat hazard alerts directly to citizens in high-risk barangays before peak afternoon sun (11:00 AM - 3:00 PM).
• ePLDT VITRO Cloud: Hosts the geospatial simulation models and powers the public-facing high-concurrency web portals during heatwave events.
• PLDT Enterprise Smart City Portfolio: Sold to progressive LGUs and educational institutions as a proactive climate-adaptation smart health module.

Demo Scenario (3 minutes for judges)

1. The Classroom Crisis (0:00 - 0:45): The presenter plays a brief video snippet of an un-airconditioned high school classroom in Dagupan City with the heat index reaching 47°C. A student faints, highlighting the delayed response of manual cancellations.
2. The 48-Hour AI Heat Warning (0:45 - 1:45): The screen reveals InitAlert PH’s interactive map for Metro Manila and Pangasinan. The AI displays a 48-hour thermal forecast: 8 schools in the city are highlighted in bright purple (“Extreme Danger: 46.2°C indoor index expected on Thursday at 1:00 PM”).
3. One-Click DepEd Advisory (1:45 - 2:30): The presenter demonstrates the Superintendent portal. With one click, the system auto-drafts the official suspension memorandum citing DepEd Order No. 037, s. 2022. It triggers a batch SMS notification to 12,000 registered parents: “Abiso mula sa LGU: Dahil sa matinding init bukas (46°C heat index), modular distance learning muna ang lahat ng mag-aaral.”
4. Outdoor Worker Safety (2:30 - 3:00): A simulated Smart SMS alert is sent to a construction foreman: “DOLE Heat Alert: Heat index has breached 43°C. Mandatory 20-minute water break every 40 minutes of outdoor labor enforced until 3:00 PM.”

30-Hour Build Plan

Suggested stack: Python (FastAPI, NumPy, Pandas, Scikit-Learn), Next.js, Tailwind CSS, Leaflet/Mapbox, Semaphore SMS API, Supabase.
Public data / APIs: PAGASA Heat Index Bulletins, Open-Meteo Hourly Forecast API, NASA POWER Solar and Meteorological Database, DepEd School Directory open data.

Risks & Privacy Notes

• Accuracy Disclaimers: Weather and heat predictions are probabilistic. The platform must explicitly cite PAGASA as the official meteorological authority while positioning itself as an advisory analytics layer.
• Privacy (Clause 24 Compliance): Parent and worker phone numbers are salted and hashed; no student grades, identity documents, or personal health records are processed.

AI Use Disclosure (draft)

InitAlert PH utilizes regression-based spatial interpolation and microclimate neural networks to downscale atmospheric temperature and humidity. Generative LLMs draft standardized DepEd suspension memos based on official administrative templates. All generated decrees require digital authorization by an authorized school administrator or LGU official.

Rubric Scoring