University of Glasgow data scientist wins the IoA Innovative Research & Application award
A University of Glasgow data scientist has picked up a prize at the first annual Analysts Awards from the Institute of Analytics (IoA). The IoA is the professional body for analytics and data science professionals. The awards were created to celebrate data’s positive impact on business and society and were judged by a panel of data analytics and education experts.
Dr Colin Torney, a senior lecturer in the School of Mathematics & Statistics, has been named as the winner in the Innovative Research & Application category for his groundbreaking work on ecological data analytics. Dr Torney was nominated by his colleague Prof Dirk Husmeier, Chair of Statistics at the University of Glasgow.
His research, which uses machine learning to investigate the movement and population dynamics of wildebeest herds in the Serengeti, will help inform decision-making around conserving and promoting wildlife in the face of human disturbance. This project is ambitious in its aims, and is built on collaboration with domain experts in the field of ecology.
While there have been animal monitoring projects in the region, this study is able to combine macro insights on animal behaviour with micro data at the level of an individual animal in migration. Special collars were designed to monitor the GPS movements of wildebeest in real time and send information on animal location and behaviour over a radio network. The data, combined with observational data from ecology experts on the ground, have made it possible to uncover data markers of key wildebeest behaviours such as grazing (from data on head tilts) or running. By tracking changes in natural behaviour of the wildebeest when they enter areas with high human activity, ecologists hope to understand the threat disturbances such as tourism pose to local animal populations better.
The team used Edge Computing to reduce the volume of data that needs to be transmitted from each animal. The collars record behaviour at a granular level, producing 50 separate readings in each of 5 categories every 10 seconds. Algorithms deployed in the devices can predict a class of behaviour: walking, lying down, standing and grazing. By carrying out some of the analysis at the Edge, it greatly reduces the volume of data that then needs to be transmitted to the central data collection points.
Other techniques were also used to support their methods, findings and conclusions, including image recognition. We already have algorithms that can pick up whether an animal in an image is likely to be a wildebeest, but getting a labelled data set to automate the counting of those animals in captured images is not easy. This is especially true for wildebeest which live in herds of several thousand up to one hundred thousand.
Dr Clare Walsh, IoA Director of Education and a member of the judging panel said: “Dr Torney has been studying the collective movement and population dynamics of animals using advanced machine learning and digital image processing to identify individual animals. While there has been work at the level of species identification in the past, the level of detail in this study, linking up properties and observations at a micro level, marks it out as a major step forward in the use of animal population studies.
“In terms of the technology used, he has had to combine different approaches to develop these solutions, working on very large data sets. The use of multiple methods of analysis that span the mathematical and machine learning spectrum shows how important it is to develop a range of techniques to solve complex problems. It is valuable research and marks a way forward for all of us.
“The work is also valuable to us around the world, as a tool to accurately measure the impact of changing human activity on animal populations. We have, in recent years, invested heavily in the idea that data can save the world from the destructive impact of environmental change, largely because we’re not sure what else to try at this point.
“For all these reasons, we selected Dr Torney as our worthy winner because he is breaking through that barrier of more detailed insights to make the best decisions.”
Dr Torney said: “I’m proud to have been chosen by the Institute of Analytics to receive this award.
“Data science offers us opportunities to see the world in new ways and can help us to make better-informed decisions across a wide range of topics.
“I’m pleased that my work has helped to support decisions about the ecology of a critically-important part of the world, and I look forward to continuing to find new ways to learn more about patterns of animal movement and behaviour in the years to come.”
The 2023 IoA Analysts Awards will be launched later this year.