From Data to Decisions: Designing a Soil-Moisture Forecasting System for Smart Irrigation
Abstract
In recent years, artificial intelligence has experienced remarkable growth. Although AI encompasses a broad range of methods and application domains, the term is increasingly used mainly to refer to large language models and agentic systems capable of performing a variety of tasks.
However, these general-purpose services are not always sufficient for addressing highly specific real-world challenges. Effective deployment in specialised domains still requires a deep understanding of the application context, carefully designed data pipelines, and machine-learning models tailored to the problem and trained on relevant data.
This talk presents the end-to-end research and development process behind a smart irrigation system. Starting from problem definition, it will cover data collection and preparation, data modelling, neural-network design, training, evaluation, and interpretation of the results.
The focus will be on a soil-moisture forecasting system that integrates in-field sensor measurements with online weather-service data. The session will discuss the main design choices, practical challenges, and lessons learned in transforming heterogeneous environmental data into actionable predictions to support more efficient and sustainable irrigation decisions.
Speakers
Marco Del Coco
Ricercatore @ CNR
Marco Del Coco is an AI researcher and engineer passionate about turning data into practical, intelligent solutions. His work sits at the intersection of machine learning, computer vision, data science, and real-world applications, with a particular interest in environmental monitoring, smart agriculture, and sustainable technologies. He combines the curiosity and methodological rigour of academic research with hands-on experience in industrial and applied projects. From designing deep-learning pipelines and data-driven services to working with heterogeneous sensor data and intelligent systems, Marco enjoys bridging the gap between research prototypes and technologies that can create measurable impact.