The Analytical Services Program develops new analytical products and provides analysis services to producers and users of social and economic data. Its major areas of focus are time series analysis (including seasonal adjustment); the construction and interpretation of socio-economic indexes; the construction of small domain estimates; the development of data integration methods; the development of data confidentialisation methodology; and general modelling and analysis of data from statistical and administrative collections. The program also provides a statistical consultancy service to other government agencies.
By providing new and innovative analytical products, the program contributes to improving the ABS’s statistical outputs and helps to inform evidence based decision making. Data confidentialisation methods, especially when applied to microdata files, allow researchers outside the ABS to construct analytical outputs of their own choice.
Time series in seasonally adjusted and trend form are an important input to policy formation, decision making and research in government, academia and the private sector. Government users include Australian government economic policy agencies (especially the Treasury and the Reserve Bank of Australia), State Treasuries, the Australian Government, state and territory governments, and local government service delivery agencies.
The main outputs of the Analytical Services Program include the ongoing seasonal adjustment of more than 2,300 time series, specialist analytical support for improving existing ABS methodology, and the development of new analytical products for potential dissemination to data users and others in the National Statistical Service. The program develops a range of analytical products including analytical frameworks, databases and models to measure socio-economic concepts. The program undertakes ongoing methodological development work to enhance data integration activities, to industrialise the generation of small domain estimates and to improve user access to information in microdata files and tabular outputs while protecting the confidentiality of individual providers. The program also develops a range of software applications for seasonal adjustment and confidentialising ABS publication tables.
The research outputs of this work are disseminated via a series of Research Papers (ABS cat. no. 1351.0). These papers may contain analytical work in preliminary stages, in order to stimulate discussion and feedback. Time series products and analyses appear in many ABS publications on a regular basis.
The main medium-term developments in the program are to:
- develop a methodology for linkage of multiple datasets - due June 2014
- undertake analyses that demonstrate the power of longitudinal datasets held by the ABS - due June 2014
- develop a methodology to analyse integrated data and dynamic confidentiality routines to support the dissemination of integrated data - due June 2015
- undertake analyses that demonstrate the power of big data held by the ABS, including administrative data and transactional data - due June 2015
- develop a methodology for confidentialising geospatial data - due June 2015
- develop methods to industrialise the generation of small domain estimates - due June 2015
- develop methods to produce error bounds for seasonally adjusted and trend estimates - due June 2016.
Ric Clarke, Assistant Statistician (A/g), Analytical Services Branch