Title: Unlocking the Power of Optimisation: Enhancing Statistical Programs at ABS

Presented by: Australian Bureau of Statistics (ABS)

Join us for a special edition OPTIMA Seminar to discover how the ABS is leveraging optimisation techniques to tackle fascinating challenges and enhance their statistical programs.

First up, we’ll dive into how we’re boosting the quality of our statistics. By cleverly addressing coverage bias and measurement errors from multi-sourced data, we’re ensuring our numbers are more accurate and reliable.
Next, we’ll examine how we safeguard sensitive information. Our practical approach to enhancing tabular outputs reduces disclosure risks, ensuring published tables maintain confidentiality without sacrificing detail.
Finally, we’ll highlight improvements in noise distribution design. By integrating a differential privacy framework, we’re making more informed decisions to better balance confidentiality protection with the delivery of meaningful insights.

Whether you’re an optimisation expert or just curious about how these techniques can transform statistics, this talk promises to be both informative and inspiring. Come see how ABS is leading the way in statistical innovation!

Speakers:
Ryan Covey joined the ABS Methodology and Data Science division in 2022. He has a PhD in econometrics from Monash University where he was also a teaching associate for mathematics and data science. Prior to that, Ryan held several roles at Telstra in data science, telecommunications and cyber security.

Chris Mann is an Assistant Director in the Data Access and Confidentiality Methods Unit (DACMU) at ABS. He started as a graduate in what is now Methodology and Data Science Division (MDSD) in 2007, and has experience in statistical methodology covering sampling, estimation, time series analysis and statistical disclosure control. As part of this work, Chris has learned just enough about constrained optimisation to be dangerous.

Joseph Chien is the Director of the Data Access and Confidentiality Methodology Unit (DACMU) at the Australian Bureau of Statistics (ABS), where he has served for over 20 years. DACMU leads the development of advanced confidentiality methods to enhance data accessibility for research while safeguarding provider privacy. Joseph was the ABS’s first Sir Roland Wilson Scholar, with a PhD focused on using administrative data to explore the microdrivers of productivity. His research interests include productivity analysis, network modelling, the semantic web, and synthetic data.

—

This event is Hybrid

• Zoom – Meeting ID 873 1557 5255; Password 778635; and
• In-person – Melbourne University
  Level 1, Room 162 “New Seminar Room” – Peter Hall Building, University of Melbourne,  813 Swanston Street (enter via Monash Rd), Parkville
  This is a short 4 min walk from Melbourne Connect (see google map)
  

SEMINAR: TUE 27 MAY 2025 11:00-12:00 (AEST, Melbourne Time)