Forget regional weather apps that tell you what’s happening 20km away. The future of sensing is hyper-local, and it’s sitting right in your garden. In this talk, we’ll explore how to transform a low-cost GNSS (Global Navigation Satellite System) receiver into a sophisticated environmental sensor capable of "feeling" the water vapour in the air before the clouds even form.

By leveraging multi-band GNSS signals, we will see that the signal delay caused by the atmosphere isn't just an error to be corrected, but a goldmine of data.

What we will cover:

- An intro into the Physics: How to use empirical models to derive the precipitable water vapour, and isolate the wet component from atmospheric noise.

- Machine Learning for Nowcasting: While empirical models interpret the present, we need ML to predict the future. We will see how to handle the non-linear dynamics of atmospheric chaos to answer the question: "Will it rain in 30 minutes?".

- Edge AI: A practical look at deploying these models on a Raspberry Pi 5 using the ONNX runtime for real-time, proactive sensing at the edge.

This is a journey through satellite signals, Python code, and the challenges of DIY hardware. Whether you are an IoT enthusiast, a data scientist, or just a nerd who wants to save the Sunday BBQ from an unannounced downpour, this talk will show you how to build a proactive system from scratch.