Live soil data + explainable AI → clear trends, honest confidence, actionable tips.
Model: Ridge regression (L2). Falls back to persistence if data is sparse. Confidence reflects recent data density.
| Date | Time | Air °C | Humidity % | Weather | Zone | pH | Soil °C | Moisture % | EC µS/cm | Notes |
|---|
Every time we measure the soil, we learn a bit more about how nature works. Three main things tell us how healthy our soil is:
By watching these three values together, we can tell when to water, when to rest, and how our soil is changing over time. That’s how science helps us take care of the planet — one small garden at a time.
Our AI model learns directly from the measurements we collect. It looks at the last few days of data and tries to predict what will happen next.
It uses things like air temperature, humidity, soil temperature, EC, and even time of day to make its guess. If we don’t have much data yet, the AI just assumes the soil will stay about the same — that’s called a persistence forecast.
The more data we collect, the smarter the predictions get. It can then suggest actions such as “Water soon” or “No irrigation needed”.
This model is called a Ridge Regression — a simple kind of math that finds gentle patterns in the data without overreacting to noise. It’s not magic, it’s mathematics helping nature 🌿.