Digital innovation enables modern life, and has the power to provide insights into solving a broad range of societal problems. Ever-increasing quantities of data, of diverse types and complexity, are being generated. We need strategies for managing and organising this data, leveraging it for predictive purposes, and exploring its impact on human society.
With a critical mass of researchers in this field, our research encompasses the full data science stack from cloud computing and databases for managing large datasets, to machine learning, data mining and natural language processing, for delivering actionable insights in applications ranging from security, privacy, social media analysis, data infrastructures, transport, enterprise systems, and health.
Case study: Big data tools have big application in cancer health risk analysis
The rising incidence of lung cancer in ‘never smokers’ with no family history of cancer has baffled Australian researchers. Of particular concern for clinicians and researchers at the Peter MacCallum Cancer Centre in Melbourne was the lack of insight into the underlying causes of these unexpected cancers – especially in the younger population. They suspected air pollution as a contributor, but could not investigate the link.
That changed in 2015, when Professor Richard Sinnott gave a talk at the centre about big-data analytic platforms that extract knowledge from diverse datasets.
Professor Sinnott is the director of eResearch at the University of Melbourne where a 20-strong team of engineers have built the Australian Urban Research Infrastructure Network (AURIN) platform and a multitude of big-data (cloud) processing systems.
Through a web-based portal, AURIN allows secure access to highly diverse social, economic and environmental data scattered in more than 70 government, academic and commercial organisations.
Professor Sinnott demonstrated AURIN’s clinical power by uploading data from the Victorian Lung Cancer Registry to look for correlations with air pollution.
In visually graphic forms, AURIN mapped the incidence of registered cancer patients against pollution levels measured by the National Pollutant Inventory, traffic data from VicRoads, the location of parks, air-quality data from the Environment Protection Authority Victoria and much more.
“We demonstrated that AURIN makes it possible to visualise disease incidence in their socioeconomic, geospatial and environmental contexts so that researchers can detect associations and connections between possible risk factors,” Professor Sinnott says.
He has since expanded AURIN’s reach of by recruiting “citizen scientists” to collect environmental data in zones associated with cancer clusters. The volunteers use mobile measuring devices (AirBeams) that are synched to the user’s smartphone through an especially designed app that shares the measurements with the central database.
The power of AURIN to make data available for clinical use is typical of Professor Sinnott’s IT solutions. Currently, a family of web-accessible, clinical IT systems support national and international research into diseases such as diabetes, cancer and a multitude of rare disorders.
“My clinical IT systems are intended to function as more than a big database,” Professor Sinnott says. “They are meant as ‘virtual research environments’ that enable far more efficient use of existing data and resources. The system we built for adrenal tumours (ENS@T-CANCER), for example, is now the go-to global resource and supports more than 25 major clinical trials.”
Research Program Leader
Professor Lars Kulik
T: +61 3 83441348