Technical Infrastructure

Behind Pribilof Traffic is a distributed network of Raspberry Pis, Docker containers, AI models, and custom Python applications collecting real-time data from the Pribilof Islands.

2
Raspberry Pi Nodes
15+
Docker Containers
3
AI Models
20+
Python Scripts

Remote Data Collection

Raspberry Pi nodes deployed across the Pribilof Islands

PAPB ADS-B Receiver

Downtown St. George

Tracks aircraft within 250+ nautical miles using ADS-B signals.

  • Raspberry Pi 4 (4GB)
  • FlightAware Pro Stick Plus
  • 1090MHz antenna
  • tar1090, readsb, fr24feed
Raspberry Pi Docker Python

PASN AIS Receiver

St. Paul Harbor

Marine traffic receiver tracking vessels in the Bering Sea.

  • Raspberry Pi 4 (4GB)
  • RTL-SDR dongle
  • 162MHz VHF antenna
  • AIS decoder
Raspberry Pi Docker

PAPB ACARS Receiver

Downtown St. George

Decodes ACARS messages - datalink communications between planes and ground.

  • Raspberry Pi 4 (4GB)
  • RTL-SDR dongle
  • 131MHz VHF antenna
  • acarshub, vdlm2dec
Raspberry Pi Docker

St. George Webcam System

Downtown St. George

IP camera with automated archival, timelapse generation, and AI analysis.

  • Network IP camera
  • Image capture every 15 seconds
  • Daily/weekly/monthly timelapses
  • AI vision analysis (Moondream)
Python AI Vision Nginx Proxy

AI & Machine Learning

Local AI models running on GPU for intelligent analysis

Weather AI Observer

Vision model analyzes webcam images to verify NOAA forecasts twice daily.

  • Moondream vision model (~1.7GB)
  • Analyzes sky, sea state, visibility
  • Tracks forecast accuracy
  • Runs at 11:55 AM & 5:55 PM AKDT
Ollama Python SQLite

Pribobs AI Commentary

Humorous AI-generated weather commentary comparing St. Paul and St. George.

  • Mistral-Nemo or Llama models
  • Compares island temperatures
  • Personality-driven responses
  • Updates multiple times daily
Ollama Python

Bush Pilot Commentary

AI analyzes approach conditions and generates pilot-style commentary.

  • Mistral or Llama models
  • Analyzes crosswinds, visibility
  • Ceiling and approach conditions
  • ILS, RNAV, and circling approaches
Ollama Python

Docker Applications

Containerized services for data collection and processing

ADS-B Stack

Complete aircraft tracking infrastructure using Docker Compose.

  • readsb: ADS-B decoder
  • tar1090: Web interface
  • fr24feed: FlightRadar24 feeder
  • planefence: Geofencing alerts
  • plane-alert: Special aircraft tracking
Docker Compose Web UI

ACARS Stack

Aircraft datalink message decoding and display.

  • acarshub: Message aggregator
  • vdlm2dec: VDL Mode 2 decoder
  • acarsdec: ACARS decoder
  • Message database storage
Docker Compose Database

Ollama AI Server

Local LLM server running multiple AI models for various tasks.

  • GPU-accelerated inference
  • Multiple models loaded
  • REST API endpoint
  • Efficient model management
Docker AI Models

Python Applications

Custom scripts for data processing, monitoring, and automation

Weather Monitoring

  • system_monitor.py
  • weather_ai_observer.py
  • pribobs_commentary.py
  • bush_pilot_commentary.py
Python 3 SQLite

Webcam Processing

  • webcam_archive.py (every 15s)
  • generate_timelapse.py
  • cleanup_old_images.py
Python 3 FFmpeg

Data Collection

  • adsb_logger.py
  • wind_watch.py
  • sensor_database.py
  • twitter_bot.py
Python 3 SQLite

Data Sources & APIs

External data feeds and integrations

Weather Data

  • NOAA API (forecasts)
  • AWC METAR (aviation weather)
  • SGXA2 Station (St. George/PAPB)
  • PADK Station (St. Paul/PASN)

Aviation Data

  • ADS-B (direct aircraft signals)
  • ACARS (aircraft datalink)
  • FAA (aircraft/airport database)

Note: We feed data TO FlightRadar24 and FlightAware, but collect our own ADS-B/ACARS directly.

Marine Data

  • AIS (vessel tracking)
  • MarineTraffic
  • NOAA Marine

Web Infrastructure

Server stack and hosting

Web Server

  • Nginx (web server & reverse proxy)
  • SSL/TLS (HTTPS encryption)
  • Proxies (webcam, AIS, ADS-B)
  • Static files (HTML/CSS/JS)
Nginx SSL

Databases

  • SQLite (sensor history, AI data)
  • JSON files (real-time data)
  • File archives (images, logs)
SQLite

Automation

  • Cron jobs (scheduled tasks)
  • Systemd services (always-on)
  • Monitoring & alerts
Cron Systemd

Standing on Shoulders

None of this would exist without the SDR Enthusiasts community

Every antenna, decoder and container on this page traces back to the SDR Enthusiasts project. Their documentation taught us how to build an ADS-B receiver on a Raspberry Pi at the edge of the Bering Sea, their Docker images (readsb, tar1090, planefence, plane-alert, acarshub, AIS-Catcher and more) are what actually run on our islands, and their Discord is where we went every time something broke at 2 AM in a gale. A huge thank you to the maintainers and the whole community.

SDR Enthusiasts Guide

The definitive ADS-B reception guide — hardware, antennas, Docker stacks and feeding. Our blueprint from day one.

sdr-enthusiasts.gitbook.io/ads-b

GitHub

Source for every container we run. Open, well-maintained and constantly improved by volunteers around the world.

github.com/sdr-enthusiasts

SDR Enthusiasts Discord

Friendly, patient and fast. The people who answered our questions about remote-site reception, weak signals and stubborn SDR dongles.

Join the SDR Enthusiasts Discord