ISSR’s emerging capabilities in advanced social data analytics techniques such as predictive risk modelling and machine learning were showcased at Professor Rhema Vaithianathan’s address to the 2019 John Western Lecture in September.
Professor Vaithianathan, who joined ISSR as Professor of Data Analytics earlier this year, is at the international forefront of research and implementation in the field of data science, and is also the Director of the Centre for Social Data Analytics (CSDA) at Auckland University of Technology. She brings extensive implementation experience from the USA and New Zealand to her role at ISSR, which will be building on her 2019 appointment by establishing a node of CSDA at ISSR and employing a postdoctoral data science researcher in 2020.
Professor Vaithianathan is internationally recognised for the end-to-end implementation of machine learning tools in high stakes government systems. This includes the Allegheny Family Screening Tool in Allegheny County, Pennsylvania, USA – an advanced algorithm to help call centre staff to triage child maltreatment allegations. This was the first application of predictive risk modelling for child welfare call screening, and an independent evaluation by Stanford University found that it reduced racial disparities in child welfare case openings and increased the accurate identification of children who needed further intervention services, without increasing the workload of investigators. Predictive risk models such as this can be used to automatically generate a risk score for individuals who are eligible for services or come into contact with an agency, providing a cost-effective way of population risk screening without additional data entry by frontline staff such as clinicians, teachers or social workers.
At the John Western Lecture, hosted by ISSR and the UQ School of Social Science at Customs House, Professor Vaithianathan outlined the important role that ISSR can play in assisting government agencies to apply advanced data analytics to complex health and social problems such as child welfare, homelessness and mental health. Her presentation Data analytics in the public sector – the tortoise or the hare? contrasted the largely cautious use of data analytics in the public sector with the comparatively quick, and profitable, uptake of data analytics seen in the private sector. She told the audience that while it may seem at first glance that the private sector is leading the way with data analytics, several recent high-profile controversies contradict that idea.
“Examples like Cambridge Analytica and the COMPAS recidivism tool in the United States demonstrate what can happen when organisations embrace data analytics technology at speed without the right guardrails in place,” she said. “We see that when it is used with no social licence, without humans at the centre, in a non-transparent way or with weak data rights, those uses can become unacceptable very quickly. The private sector took to AI and machine learning early and fast, seeing data as a strategic asset, whereas the public sector has tended to move slower, conscious of concerns like consent, transparency and having a ‘human in the loop’.”
In her lecture, Professor Vaithianathan identified six elements of a successful data analytics strategy for the public sector: agency leadership, transparency and fairness, a community voice, a multi-disciplinary team, and ethical review and independent evaluation. She believes the future lies in the public sector leading the way by adopting human centred data science, and she looks forward to continuing conversations generated by her highly-successful John Western Lecture, which was attending by more than 160 people and resulted in numerous follow-up enquiries from key Queensland agencies.
The John Western Memorial Lecture is held in honour of former UQ Professor of Sociology John Western AM – the first professor of sociology at UQ and Head of the Department of Anthropology and Sociology in 1970.
The slides from Professor Vaithianathan’s John Western Lecture can be viewed here.
You can watch a video of the full lecture here.
Find out more about Predictive Risk Modelling Methods here.
See ISSR’s latest training course for Public Service Senior Managers to explore the use of data science to address key public policy and practice questions.
