UNIMELB_NGUYEN-Josh_VYT-LOCAL-2023.mp4

This presentation is an overview of my PhD research on the prediction modelling of future suicide risks, which I have created as a part of the 2023 Visualise Your Thesis Competition. 

Transcript: 

Let’s imagine your life is like this balloon. It fills up with pressures of everyday life, work, relationships, health and finances. But just like this balloon, everyone has a limit to how much pressure they can hold. And for some people, this pressure could become a bit too much, and leading them to consider “SUICIDE”. 

If we are able to predict when a person reaches their breaking point, what if we could intervene in time? 

That’s exactly what my PhD research aims to do. We use various data resources from health history, genetic markers, daily wood fluctuations, and social support, to even data from smartwatches that monitor activity, sleep and locations of everyday life. Then a machine learning model will be used to detect patterns from these sources and predict future risks of suicide. 

You can think of this method like a weather forecast, which draws on a massive amount of data about the regions to predict future weather using complex mathematical models. But the difference between weather forecast and suicide prediction is that we cannot change whether it will rain tomorrow. Yet we can predict who's at higher risk of suicide and provide extra support and early interventions