Here's just a small sample of projects we've worked on.

Improving prediction of Tropical cyclones

We have experience helping scientists at the Bureau of Meteorology, Melbourne to improve accuracy of tropical cyclones forecasts. Our algorithms take into account wind speeds and other environmental variables to reduce the bias in the forecast.


prediction of extreme events and risk assessment

We have developed new prediction methods to estimate the chance of extreme events taking place in space and time. We applied these methods to assess risk and inform portfolio decisions in finance. We are currently applying our techniques in the context of geo-spatial data (solar data, wind speed, etc.) and environmental risk assessment.


finding genetic factors related to cancer

We developed a number of statistics and machine learning methods able to find genetic factors associated with disease. We successfully applied our methods in the context of case-control data on breast, ovarian, leukemia and melanoma cancers.

predictive modeling of cancer cells from microscope Image data

We have experience developing new prediction methods and related algorithms able to predict growth rate of cancer cells based on high-content microscope image data. We applied our methodology on real experimental data in collaboration with the Pathology Department at the University of Melbourne to investigate how interactions between cancer and healthy cells affect their growth.

prediction of crime occurrence in residential buildings

After designing a survey and collecting data in residential buildings, we developed a crime risk assessment model. We were able to quantify the crime risk and a given building and assign a risk rating, given some basic information on the building such as geographic location, building size, number of entrances, and socioeconomic data.

analysis of cancer patient treatments with a simple blood test

We have leveraged the spatial and temporal genetic data from patients with leukemia to help guide clinical treatment and improve patient outcome. Using molecular biomarkers from a noninvasive blood test, we can rebuild the genetic landscape of tumors and identify what is causing a particular cancer and devise ways to fight it. By doing this through time we can see resistance building and change treatment to help patients live longer.