This is not football (soccer)-related like other projects in the same category, but it is still related to a sport that has football in its name and it is also data-focused, so I decided to put this project into the same category.

While undertaking my study at QUT, I had the opportunity to take on a summer research project that focused on developing a machine learning model to predict injuries in AFL athletes using GPS-collected data. The project was also supervised by Associate Professor Kerrie Mengersen, who was QUT’s Director of Centre for Data Science.

Having done most of my data work on Python and on football/soccer, it was a breathe of fresh air to shift my focus onto a different sport and worked extensively on R throughout this project. This project also gave me the opportunity to do a lot of research on machine learning and how past studies have used algorithms and models to predict injuries in the AFL and other sports.

In the end, the results that I found from the study align with past studies and my own hypotheses about injuries. If an athlete accelerates, decelerates, or changes speed more often during a long and extended training or game session, the likelihood of sustaining a high severity injury increases significantly. This leads to a lot of pressure being applied onto the athlete’s lower limb, especially on the knees.