A Collaborative Approach to Agile statistical Processing Architecture ABS Experience and Aspirations Meeting on the Management of Statistical Information Systems (MSIS 2010) 26-29 April 2010 Daejeon,
Download ReportTranscript A Collaborative Approach to Agile statistical Processing Architecture ABS Experience and Aspirations Meeting on the Management of Statistical Information Systems (MSIS 2010) 26-29 April 2010 Daejeon,
A Collaborative Approach to Agile statistical Processing Architecture ABS Experience and Aspirations Meeting on the Management of Statistical Information Systems (MSIS 2010) 26-29 April 2010 Daejeon, Republic of Korea Brian Studman IT Director, information Management Transformation Program Topics • Learning from Experience - the threads of process, metadata, architecture and collaboration • ABS revised Architecture and Collaboration • Aspiration - The Information Management Transformation Program • < 1980s – mainframe, subset of metadata used consistently, 'generalised' software lessons • 1980-89 – PCs & client-server innovation chaos • 1990-99 - standard process & process control, output warehouse (specialising output & dissemination processing), Web site constant innovation, KM environment • 2000-2009 – data capture specialisation (BSIP), ISHS2, external view through NSS & collaboration Architecture Lessons • Lack of low-level content • 'Posters' become dated too easily • Business ownership of architecture flagged • Architecture governance did not influence early enough in project • Translation from architecture into high-level design didn't consistently occur Need more in this area Collaboration Lessons • Mutual value proposition • Multi-language issue – plan/design • 1+ solution is fine • • Common standards and architecture Senior executive 'champions' • Incremental delivery – agile process • Genuine commitment • IP, licencing, ongoing support, • technology & project issues – plan to • manage and manage • • 'Lead' organisation for projects Mutual respect & attention to detail Out-posting staff Building New Business Needs e.g. Concepts, papers, ideas, standards, designs, architecture Offers to Share Existing e.g. Blaise, PC-Axis, .stat Sharing Highest Value Region e.g. Blaise & PC-Axis extensions Collaboration IMTP • About business • agility for complex information solutions • Usual business drivers: efficiency • Enabled by whole-ofGSBPM metadata • management (DDI+SDMX) • • Plus BPMS+SOA+... • • Practically further international collaboration • Heavy emphasis on metadata standards, 'active' metadata Paradata separated as a business value element Aligning with other international work e.g. SAB & CORA Focussing on NSI operations Merging: Process, metadata, architecture & collaboration SDMX DDI A g g r e g ate d d a ta Micr odata Indicator s, Time Ser ies Acr oss time Acr oss geogr aphy Open Access Easy to use Low level obser vations Single time per iod Single geogr aphy Controlled access Exper t Audience From Process to Workflow Pragmatic (ABS internal) IMTP DE R C LA M S 2 Q DT 1 Specify Needs 2 Design 1.1 D etermine needs for information 2.1 D esign outputs 1.2 Consult & confirm needs 2.2 D esign variable descriptions 1.3 Establish output 2.3 D esign data collection objectives methodology 1.4 Identify concepts 1.5 Check data availability 1.6 Prepare business case Quality Management ID W /C P C F-M at 4 3 Build 5 / M etadata M anagement 4 Collect 5 Process 4.1 Select sample 5.1 Integrate data 3.2 Build or enhance process components 4.2 Set up collection 5.2 Classify & code 3.3 Configure workflows 4.3 Run collection 5.3 Review, validate & edit B laise E D R 4 Build3.13.1data 5 Build data collection collection instr ument instrument A 1 7.17.1 Update B Update output output S systems systems 9 8 I W 1 7.2 7.2 6.2 6.2 S Produce Produce Validate Validate dissemination W dissemination outputs outputs products 7 A 2 products 3 T D D7.3I S tds 7.3 7.3 6.3 1 MMManage anage anage Scrutinize & release of release releaseofof explain 2 dissemination 1 O E C D . s t a t dissemination dissemination products products products RE E M 2 3.4 Test production system 2.5 D esign statistical processing 3.5 Test statistical business 5.5 D erive new variables & methodology process statistical units systems & workflow 7 Disseminate 6.1 6.1 Prepare Prepare draft draft outputs outputs 6 2.4 D esign frame & sample methodology 2.6 D esign production 6 Analyse 4.4 Finalize collection 5.4 5.4 Impute Impute 2 7 6.4 6.4 Apply A pply disclosure disclosure control control 6.5 Finalize outputs 5.6 5.6 Calculate weights Calculate weights 5.7 5.7 Calculate Calculate aggregates aggregates 5.8 Finalize data files support 9.2 Conduct evaluation repository 8.3 Preserve data and 9.3 Agree Action associated metadata plan 1 T re a s u ry 1 DSIP strands Strand 1 – Leverage off SDMX capabilities provided by OECD.stat Strand 2 - REEM Strand 3 – Publishing standards using DDI Strand 4 – QDT->DDI->Blaise; HSF1(Blaise)->DDI->HSF2(QDT) 5 Strand 5 – IDW/CPCF Mat->DDI<->EDR Strand 6 - Administrative data acquisition 6 < fo o ter> 8.2 M anage archive 4 7 8 < d ate/tim e> 9.1 Gather evaluation inputs CS DI 7.5 M anage user 3 7 8.1 D efine archive rules associated metadata 2 7 system 9 Evaluate 8.4 D ispose of data & 7.4 Promote dissemination products 1 3.6 Finalize production 8 Archive 9 Strand 7 - Estimation/Imputation/Macro editing Strand 8 – Time Series adjustment – Replace SEASABS Strand 9 – National Accounts < n u m b er> A Collaborative Approach to Agile statistical Processing Architecture ABS Experience and Aspirations THANK YOU Brian Studman IT Director, information Management Transformation Program