Chris McKendry Australian Bureau of Statistics ABS-SSAI Careers Night 12 April 2012 My Background Bachelor of Science at UNSW Honours in Statistics Thesis on Time Series ABS Cadetship Work Experience in Time Series between semesters Study support for Honours Graduate Program (Jan 2011) 13 months in Time Series Recently moved to Survey Methodology Perks I Experienced at the ABS Consistent extra-curricular activities Trivia, tennis, soccer, basketball, touch footy, board games Access to training SAS, Project Management, Economic Statistics, Statistical Reporting, Time Series, Survey Methods Networking and Travel Interstate travel is completely normal Represent your section / division / organisation in various forums / events Perks of the ABS FLEXITIME (!!!!) Like to start late and finish late??? Like to start early and finish early??? Wish to coordinate work hours with other studies??? Have kids to pick up from school??? Like to combine public holidays with flexi-hours??? Colleagues with similar talents and interests Programming, Maths, Statistics, Econometrics Graduate Year in MDMD Rotations Scheme Rotate at 10 months and 18 months Develop >= 3 disciplines in the first 2.5 years Sections of work Data Confidentialisation, Data-Linking, Operation Research, Sampling, CPI estimation, Time Series, Econometrics Corporate Responsibilities Coordinating social activities Participate in MDMD seminars Projects assigned to 2011 Grads Friend Project Relevance Joseph Data confidentialisation Researcher access to confidentialised unit records Aparjit Monthly Population Survey National estimates of labour statistics Daniel Optimality of the CPI Integrity of national statistics Carrie Data linking Joint use of admin and census data Me Time Series Reconciliation Detailed data for socio-economic analysis My Grad Project Is a pretty good example of an output useful to the business / policy / research community Not meant to be Earth shattering But demonstrates what kind of technical skills are useful within MDMD My Grad Project Background in Seasonal Adjustment Monthly Hours Worked In Australia (Original) Original Data • Seasonality • Moving Public Holidays • Moving School Holidays My Grad Project Background in Seasonal Adjustment Monthly Hours Worked In Australia (Seasonally Adjusted) Seasonally Adjusted • Less volatile • Comparable Months • Complimentary to Unemployment Rate My Grad Project Background to ABS Hours Worked Series Relevance Users • Key economic indicator • RBA • Effects of Policy • Treasury • Natural Disasters • ABS • Economic Trends / Shocks • DEEWR Cross-Sectional Splits Sex State Part/Full time Age My Grad Project Reconciliation Publish SA HW in Sex by Status splits BUT • Seasonally adjustment not linear • Sum of SA Components ≠ SA Adjusted Aggregate Sex Male Job Status Female Person Full Time ℎ, ℎ, ℎ, Part Time ℎ, ℎ, ℎ, Working ℎ, ℎ, ℎ, My Grad Project Reconciliation Solution • Time series reconciliation • 6 constraints to meet for each time (easy and standard) Sex Male Job Status Female Person Full Time ℎ, ℎ, ℎ, Part Time ℎ, ℎ, ℎ, Working ℎ, ℎ, ℎ, My Grad Project Reconciliation: 3D Aggregation Structure 3D Aggregation Structure • 63 constraints PER OBSERVATION Sex • Sex by Status by State = 1, … … , 111 ……… ⋮ ⋮ ⋮ 11 ⋮ ⋮ ⋮ 11 ……… 1 = 1, … … , Quality Concerns • Meet constraints • Maintain Level • Maintain Temporal Movements F/P- time 11 ……… ⋮ ⋮ ⋮ 1 1 ⋮ ⋮ ⋮ ……… 1 My Grad Project You wanted maths at work right?