opensensor.space

Last refreshed 1 min ago

opensensor.space cloud-native architecture

Quick Station Access

Near Real-Time Dashboard

View the latest 15-minute weather readings

Environmental Monitoring Dashboards

Temperature and Humidity

Temperature trends and humidity analysis

Gas Sensors

Oxidised, reducing, and NH3 measurements

Light and Proximity

Ambient light levels and proximity detection

Atmospheric Pressure

Barometric pressure trends and analysis

Particulate Matter

PM1.0, PM2.5, and PM10 air quality analysis

System Health

CPU, memory, disk, and network status

From Traditional IoT to Cloud-Native

Traditional IoT relies on local hubs - sensors sending data via MQTT to central servers running databases like InfluxDB. This introduces complexity and potential data loss during outages.

Our First Trial

In early 2024, we built our first environmental monitoring station - a Raspberry Pi Zero W with an Enviro+ sensor connected to an Intel NUC hub running TimeScaleDB. It worked, but when the hub disconnected, all readings sent via MQTT were lost.

After a year of experimenting, we had a realization: Why use extra energy and resources when the Raspberry Pi already has WiFi?

This insight became the foundation of opensensor.space - eliminating unnecessary infrastructure by leveraging what edge devices already have. Simplicity scales better than complexity.

opensensor.space cloud-native architecture diagram

Architecture Overview

The pattern works for any IoT data source. Explore the data structure using DuckDB:

SUMMARIZE
    SELECT
        *
    FROM
        read_parquet('s3://us-west-2.opendata.source.coop/walkthru-earth/opensensor-space/enviroplus/station=019ab390-f291-7a30-bca8-381286e4c2aa/**/*.parquet', union_by_name=true, hive_partitioning=true);

Sensor Data Statistics

Min, max, average, and other key metrics for all sensor readings
No Results

Edge devices stream measurements directly to cloud storage in Parquet format - no databases required.

How It Works

  1. Sensors collect readings every 5 seconds, batched into 15-minute Parquet files
  2. Files stored in Hive-partitioned format: station={ID}/year={year}/month={month}/day={day}/data_{time}.parquet
  3. Data queried directly using DuckDB-wasm in the browser with Evidence.dev

Our data is hosted on Source Cooperative - free S3-compatible storage for open datasets (requires approval from Radiant Earth). Any S3-compatible storage works.

Example: Environmental Monitoring

Reference deployment using Enviro+ sensor: temperature, pressure, humidity, gas sensors, light, proximity, and particulate matter (PM1.0, PM2.5, PM10).

Platform Benefits

  • Low carbon footprint - Edge processing reduces data transmission by 60-90%
  • Cost effective - No databases, message brokers, or backend servers
  • Scalable - Object storage scales from single sensors to millions
  • Hardware agnostic - Works with any Python-capable device (Raspberry Pi, NVIDIA Jetson, ASUS Tinker Board, etc.)
  • Resilient - Offline-first with automatic sync when connectivity returns
  • Open - Standard formats (Parquet, S3) accessible to any analytics tool

Sensors operate autonomously even when offline - zero data loss during network outages.

Roadmap

  • GPS-enabled sensors - Location tracking for mobile sensor networks
  • Multi-sensor support - LoRa, radio (AIS), industrial sensors, agricultural IoT, smart city applications
  • One-line setup - ✅ Completed (see Join the Network)
  • Multi-cloud support - ✅ Completed (AWS, GCS, Azure, Cloudflare R2, Wasabi, Hetzner, MinIO)
  • Stations health - ✅ Completed (CPU, Memory, Disk, WiFi, NTP monitoring)

Resources

opensensor.space is an initiative by walkthru.earth