Times: 
5:00pm Pre-drinks and nibbles in foyer, ABS House, 45 Benjamin Way, Belconnen (Map)
5.15pm SSA AGM in Knibbs lecture theatre
6.00pm Presentation in Knibbs lecture theatre
7.30pm After the talk, there will be dinner at Bella Vista, 84 Emu Bank, Belconnen (Restaurant)

Please RSVP Warren Muller (Warren Muller or reply to this email directly) by Saturday 25 August if you would like to attend the dinner. 

Speaker: Daniel Elazar, ABS

Topic: Methodological Challenges in Balancing Data Confidentiality and Utility in an Increasingly Contested Data Environment

National Statistical Offices (NSOs) are tasked with maintaining the appropriate balance between upholding legislative obligations relating to the privacy and confidentiality of individuals and businesses while maximising the utility of data releases for policy makers and researchers. In Australia, the Data Integration Partnership for Australia (DIPA) is a major investment to maximize the use and value of the Government’s data assets while preserving the privacy of individuals which has led to a rapid growth in demand from policy makers and researchers for more accurate and detailed longitudinal data. However, there are also increasing challenges for NSOs in ensuring that data releases are sufficiently protected in a world witnessing massive growth in the availability, size and diversity of data sources (including social media, sensor and transaction data), growing sophistication of privacy attacks, availability of large-scale computational power and importantly, growing concerns over personal privacy. 

This talk will discuss the evolution of and statistical methodologies underpinning ABS TableBuilder for the automated and interactive confidentialisation of tabular outputs and examine whether certain aspects of differentially privacy could be leveraged.

Bio: 
Daniel Elazar has worked in Methodology Division at the Australian Bureau of Statistics for slightly over thirty years. He is currently the Director of the Data Integration, Access and Confidentiality Methodology Unit which develops and applies statistical methodologies in record linkage and confidentiality. Previously Daniel led the team developing the methodologies underpinning the Data Analyser prototype system for confidentialisation of analytical outputs produced from unit record data and has also been involved in the methodological development and maintenance of the TableBuilder product. He also work extensively in the areas of sample survey methodologies, small area estimation and the use of satellite sensor data for crop estimation. In his spare time, Daniel enjoys spending time with his family, peripatetic activities (going for walks), catching up with friends, gardening and the odd movie.

Daniel has an undergraduate degree in mathematics from University of New South Wales and a postgraduate diploma in statistics from the Australian National University.