Social services agencies around the world face incredibly complex, high stakes questions. How do they fairly distribute limited resources? Identify those most in need? Understand risk levels and take steps to prevent harm?
Researchers at the AUT Centre for Social Data Analytics (CSDA) are helping agencies use data to find answers and make better decisions about how to support the most vulnerable people in their communities.
CSDA founder and director Professor Rhema Vaithianathan says the centre’s research team of data scientists, economists and software engineers builds computational models and assessment tools to process the vast quantities of data available to government and social service agencies.
Looking into the data helps agencies to better understand their own practices and to direct services to those at highest risk for adverse outcomes.
“Our predictive risk modelling provides rapid insights about risk levels for individuals in the longer term, giving staff in these agencies additional information to help them make consistent, quality decisions. The volume and complexity of the available data means agency staff couldn’t easily analyse it or make use of it in their decision-making without these models.”
One of the agencies CSDA has been working with since 2014 is Allegheny County’s Department of Human Services (DHS) in Pennsylvania, US. CSDA models are already successfully in use by DHS staff to support call-screening decisions about children reported as being subjected to maltreatment and neglect.
New models are now helping Allegheny County prioritise harm prevention services for newborns and their families through the Hello Baby programme, and to triage requests for housing assistance so that people at greatest risk of harm associated with homelessness get help first.
CSDA has developed a global reputation for its research and social impact and Professor Vaithianathan says the focus is now on working with collaborators who can help scale out the social impact of the centre’s work.
International research partners include the Children’s Data Network at the University of Southern California, the Institute for Social Science Research at The University of Queensland, which hosts a second CSDA site, and scale-out collaborators Mathematica in the US.