How can researchers use data and machine learning to address some of the world’s most pressing social issues, while actively engaging with end users and impacted communities? UQ’s Institute for Social Science Research presents a workshop with Professor Rhema Vaithianathan, Professor of Social Data and Analytics at ISSR and Director of the Centre for Social Data Analytics (CSDA) at Auckland University of Technology (AUT) in New Zealand. 

Professor Vaithianathan will discuss work on a variety of international projects that use machine learning for social impact. Recent projects include the development and implementation of child welfare decision-support tools including the Allegheny Family Screening Tool (which has been featured in a range of media outlets including The New York Times), a project with the Chilean Government to use linked administrative data to protect families, RCTs of decision-support tools in Colorado with social workers, and a homelessness decision support tool for Allegheny County.

Following Rhema, Diana Benavides Prado (AUT) will discuss some of the challenges to deployment of machine learning in real-world, high-stakes settings and PhD students Ola Zytek (MIT) and Chris Mills (Princeton) will present work they are doing with Rhema on providing Larimer  County, Colorado (USA) with an explainable decision support tool for frontline child protection workers. Gayani Tennakoon (ISSR) will share her research investigating whether Queensland pediatric intensive care data can be used to help clinicians identify patients at risk of poor long term educational outcomes.

Webinar Recording


Professor Rhema Vaithianathan

Rhema Vaithianathan is a Professor of Social Data Analytics at the Institute for Social Science Research, University of Queensland. She is also a Professor of Economics at Auckland University of Technology (New Zealand) where she is the founder and co-director of the Centre for Social Data Analytics.

Rhema is recognised internationally for implementation of machine learning tools in high stakes government systems such as child welfare. She leads the international research teams that developed, and continues to refine two currently implemented child welfare predictive risk modeling tools: the Allegheny Family Screening Tool (Allegheny County, PA) and the Douglas County Decision Aide (Douglas County, CO)  Rhema’s other predictive analytics work in the United States is diverse, including leading development of a PRM tool to help Allegheny County better prioritize homeless services. Her work has been published in top journals and profiled in media outlets including The New York Times and Nature News. She is frequently invited to speak to government agencies, researchers and practitioners around the world about ethical use of machine learning tools in public policy. Rhema has held research positions in Australia, Singapore and the United States, including a Harkness Fellowship at Harvard University.

Dr Diana Benavides Prado

Diana completed her Bachelor of Systems Engineering in 2010, and Master’s in Systems Engineering and Computing Sciences in 2012, both in her native country, Colombia. From 2010 to 2016 she worked as a researcher in science and technology for a variety of Colombian government institutions in sectors such as finance, healthcare, justice and geosciences. Diana moved to New Zealand in 2016. Since then she has been working as a Research Fellow in Data Science for the Centre of Social Data Analytics, AUT. Her experience spans a variety of projects in data science and machine learning, as well as teaching and tutoring in algorithms, programming and machine learning. Her research areas are both in fundamentals of machine learning in areas such as transfer learning, lifelong learning, and human-algorithm collaboration, as well as applied data science for decision making support in high-stakes problems. She holds a PhD in Computer Science from the School of Computer Science at The University of Auckland, New Zealand.

Ola Zytek

Ola is a PhD student with the Data to AI lab at MIT. She works on explaining machine learning predictions to end-users, to allow for smoother interactions between humans and computers. She is currently developing a tool for the domain of child welfare screening, which aims to bolster trust in a decision aide model and guide social workers though the model's predictions..

Chris Mills

Chris Mills is a Ph.D. candidate in economics at Princeton University. He is passionate about using linked administrative data to study the causal effects of child welfare interventions on child and family wellbeing, and his research interests include human capital formation in foster care, foster parent labor supply, and expert decision-making. Chris holds a B.A. in economics and computer science from Cornell University.

Dr Gayani Tennakoon

Gayani is working in the social data analytics team led by professor Rhema Vaithiyanathan. Gayani’s work focuses on developing interpretable machine learning models to help decision making with health and administrative data. She has recently involved in a research on exploring the usefulness of machine learning in predicting the educational outcome of children who admitted to paediatric intensive care units.

Gayani has completed her PhD from Queensland University of Technology in 2020. During her PhD, Gayani has worked with unsupervised machine learning methods and conducted research on discovering knowledge from social media based on user interaction patterns.



9:00am - 9:05am

Professor Mark Western, Director, ISSR

  • Welcome & introductions
9:05am - 9:20am

Professor Rhema Vaithianathan, UQ and AUT

  • Ethical Data Science for Social Impact
9:20am - 9:30am

Dr Diana Benavides Prado, Senior Research Fellow, AUT

  • Data Science and Deployment Challenges while Supporting Decision Making in Human Services
9:30am - 9:40am

Dr Gayani Tennakoon, Postdoctoral Research Fellow in Data Science, ISSR

  • Exploring the usefulness of machine learning in predicting future educational outcome of critically ill children
9:40am - 9:50am

Ola Zytek, PhD Student, MIT

  • Sibyl: Machine Learning Explanation Tool for Better Decision Making
9:50am - 10:00am

Chris Mills, PhD Student, Princeton

  • More than a Score? The Effect of Algorithmic Tools on Child Welfare Decisions and Outcomes
10:00am - 10:30am Questions & Discussion

Webinar details 

Date: Friday, 11 September 2020

Time: 9:00am - 10:30am AEST 


This webinar is organised by the Institute for Social Science Research, as part of Social Sciences Week.


About Social Sciences Week 2020

As part of Social Sciences Week, The University of Queensland's School of Social Science and Institute for Social Science Research invite you to virtually attend a series of UQ events offering insight into the impact of social sciences on our lives. 

We look forward to seeing you online.