Research Insights

Farming is a data-rich business. Everyday decisions, how much fertiliser to apply, when to treat a field, how to manage soil health — depend on reliable information. But datasets collected across multiple sites and over time inevitably contain outliers: data points that don't fit the expected pattern, whether due to measurement error, equipment variation, or genuine natural anomalies.

At Rothamsted Research, one of the world's longest-established agricultural research centres, Analisa's work focuses on improving how these outliers are detected and handled. The research draws on data collected at Rothamsted's experimental platforms in Hertfordshire and at the North Wyke site in Devon.

A poorly handled outlier can seriously skew analysis. A single anomalously low soil pH reading might suggest an entire area needs acidification treatment when it doesn't. An outlier nitrogen reading could result in over- or under-application of fertiliser, with knock-on effects for both costs and the environment. By developing better methods for identifying and managing these irregular data points, the research aims to help farmers make more accurate, evidence-based decisions, reducing unnecessary resource use and supporting more sustainable farming practices.

The longer-term potential includes informing the development of software tools and improved management strategies that could benefit the broader farming community.

Public Engagement

This animation was commissioned to communicate Analisa's research to a general audience, making the concepts of data analysis and statistical outliers accessible without requiring any prior scientific knowledge. It is intended for use in public engagement settings, conferences, and educational contexts.

This is one of a series of animations produced for Envision PhD researchers. Envision runs a competition in which early-career researchers submit a brief for an animated film; selected winners have their research brought to life through professional animation.