Knowing how much water is in the soil is the difference between watering too much, too little, or just right. For my BSc research in Soil Science, I wanted to build a low-cost soil moisture sensor that could be calibrated for local soil types — and give farmers and researchers a simple way to read it.

The hardware

The heart of the project was an Arduino with a capacitive soil moisture sensor. Capacitive sensors are a good choice here because they measure the dielectric constant of the soil (which changes with water content) without corroding the way cheap resistive probes do.

The sensor reads a raw value, but raw values alone aren't enough. Different soils — sandy, clay, loamy — hold water very differently, so the same raw reading can mean "bone dry" in one soil and "waterlogged" in another. That's where calibration comes in.

Calibrating for local soils

I calibrated the sensor against known moisture levels for several local soil types, collecting paired measurements and fitting a relationship between the raw reading and actual water content. That step turned a generic sensor into something that gives meaningful numbers in the field.

The companion app

Data is only useful if people can get to it. I built a React Native mobile app (using Expo) that connects to the sensor and displays moisture readings, so a user doesn't need a laptop to check the soil. The app was built to be simple and fast to use outdoors.

What I took away

• Calibration is everything — a sensor without context is just a number.
• Pairing hardware with a friendly app turns a research gadget into a practical tool.
• Cheap components can do serious work if you put the effort into validating them.

This project was my first real end-to-end build that crossed hardware, data and software — and it shaped a lot of what I work on today.