How P.U.L.S.E. Works

P.U.L.S.E. (Philippines Unified Live Seismic Explorer) is an automated, real-time crisis monitoring service. Below is a detailed overview of our system telemetry, data latency, and geological analysis methodologies to maintain scientific transparency and algorithmic trust (E-E-A-T).

1. Telemetry Data Feeds

P.U.L.S.E. actively aggregates raw data from multiple independent global and national agencies:

Agency Source	Refresh Frequency	Typical Latency	Purpose
PHIVOLCS	Every 15 Seconds	< 1 minute	Primary Philippine authoritative provider. Contains intensities and fault locations.
USGS	Every 1 Minute	2 - 5 minutes	Global tectonic tracking and secondary magnitude verification.
EMSC	Every 1 Minute	2 - 5 minutes	European-Mediterranean Seismological Centre event triangulation.
WeatherAPI	Coordinates-triggered	Instant on-event	Precipitation data to assess landslide hazard risks.

2. Data Processing Pipeline

When a seismic sensor registers an event, the following steps occur in seconds:

• Deduplication Engine: Coordinates are cross-referenced (1.35° spatial / 30-minute time window) to merge duplicate event alerts from USGS and PHIVOLCS, prioritizing PHIVOLCS as the authoritative local source.
• Tectonic Mappings: Raw epicenter coordinates are compared against active fault vectors (GEM Tectonic Faults Database) and subduction trench vectors (Peter Bird Trench Model) using spherical trigonometry.
• Landslide Assessment: Recent 48h precipitation maps are queried for the epicenter. If magnitude thresholds are exceeded, impact zones are mapped to identify landslide vulnerability.
• Real-time Broadcast: P.U.L.S.E. pushes verified payload updates instantly to the frontend using WebSockets (`/ws`) and updates search sitemaps via IndexNow APIs.

3. Database & System Resiliency

Our infrastructure operates on a Redis-First Hybrid strategy. The active cache is stored in high-performance Redis Sorted Sets (timestamp-scored). Persistent backups are stored in daily JSONL logs (`history_YYYY.jsonl`) ensuring zero single points of data failure.