Who’s Hanging Out with Whom? Tracking Cow Friendships and analysing variations in the social role over time

Who’s Hanging Out with Whom? Tracking Cow Friendships and analysing variations in the social role over time

Aim

Understand social networks in dairy cattle and explore how the individuals' positions change over time. We will explore this by analysing how the position in the social network and contact frequency within social networks of cows vary with time and physiological status.

Background

Understanding the social dynamics of livestock is increasingly recognized as a key element in improving animal welfare and productivity. In cattle, social contact patterns may shift over time depending on age, physiological status, such as reproductive stage or health. This project investigates how the structure of cow-to-cow social networks evolves and how individual physiological states affect social positioning in the networks and their interactions.

Social interactions among herd animals are not random; they form dynamic networks that reflect hierarchies, preferences, and behavioural states. Advances in sensor technologies (e.g., RFID, accelerometers) now enable precise, continuous monitoring of individual animals’ positions and interactions. In this case, we will use GPS-indoor data collected over the years to evaluate the changes that individuals experience over time.

Social network analyses provide a way to quantify these relationships, often using graph-theoretical metrics to describe how ‘connected’ or ‘influential’ an individual is within the herd. Fluctuations in these metrics can be biologically meaningful—for example, a cow may become socially isolated with age, during illness, or increase her contact rate during estrus.

Project description

This is a data-driven project involving:

• Cleaning and analysis of longitudinal data from sensor-equipped cows
• Mapping and visualisation of contact networks over time
• Statistical modelling of the relationship between social metrics.
• Depending on interest and background, students may also learn and apply dynamic network modelling.

Learning Outcomes

• Gain practical skills in R programming, with applications in data cleaning, visualization, and statistical modelling
• Mathematical skills in network analysis, and time-series modelling.
• Experience linking behavioural ecology with animal physiology
• Insights into real-world applications for smart farming technologies

Specifications

Experience with R will be beneficial, but not mandatory as the candidate will use a custom-developed R package for data manipulation and analysis and receive full guidance throughout the data analysis process.

Suitable for students in animal science or veterinary medicine.