SIParCS 2025 - Isaac Oppong-Baah

Microservice-Driven IoT Architecture for Atmospheric Sensornets and Real-Time Visualization

Recorded Talk

As part of the project titled “Community-Driven LoRa Deployment, Open Source Integration, and Machine Learning Evaluation for Atmospheric Sensornets,” the focus centers on designing a scalable, microservice-based architecture to support real-time environmental data collection, storage, and visualization.

The system architecture includes the configuration of Raspberry Pi edge devices and the deployment of an MQTT broker to ingest sensor readings from existing NCAR LoRa-enabled stations. These readings are processed through a designed message orchestration service and transmitted to a PostgreSQL database via a RESTful API gateway designed for efficient and modular communication.

For data visualization and exploration, Metabase was selected due to its open-source nature, ease of integration, and intuitive user interface, which reduces the technical barrier for both internal stakeholders and external station collaborators. Supporting services were developed to automate the creation of users, dashboards, collections and models, enabling rapid onboarding and consistent visualization standards across the network. This integration empowers users to gain immediate insights from sensor data without requiring advanced programming or database expertise.

Ongoing development includes the incorporation of NCAR’s Miles CREDIT weather forecast model, which will generate daily predictions and present them alongside observed measurements to support comparative analysis.

By combining real-time IoT data streams with user-friendly visualization and automated forecast integration, this framework enhances NCAR’s ability to support scalable, community-driven atmospheric monitoring and facilitates broader scientific engagement with the data.

Mentors: Agbeli Ameko, Keith Maull, John Schreck

Slides and poster