Statistical process as a structured chain of successive actions and intermediate products, supported by the coherent use of metadata by Statistics Netherlands
Download ReportTranscript Statistical process as a structured chain of successive actions and intermediate products, supported by the coherent use of metadata by Statistics Netherlands
Statistical process as a structured chain of successive actions and intermediate products, supported by the coherent use of metadata Focused on energy statistics and relation IRES Hans Pouwelse Statistics Netherlands 1 content • Brief description Business Architecture statistical process • Role of coherent meta data • Focused on Energy Statistics • Relation to IRES 2 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S Output sphere data Meta servers for P r o c e s s M e t a d a ta 3 Conceptual metadata Output world • Definitions variables and classifications in output terms (‘language’ of users) = definitions rows en colums output tables • Reporting period (year, month…) • Units of measurement • Statistical units (micro level output register) 4 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S Output sphere data Meta servers for P r o c e s s M e t a d a ta 5 Conceptual metadata Input world Questionnaires • Design questionnaires: questions, definitions (in the ‘language’ of respondents) • Reporting period • Observation units (observable units) Registrations (‘administrative data’) • Definitions of variables and type of units 6 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S Output sphere data Meta servers for P r o c e s s M e t a d a ta 7 Process metadata To provide methods and rules for the process to go from stage to stage (from database to database) • Sampling scemes • Methods and rules for editing, validation, imputation, aggregation and disclosure control • Transformation rules to bridge the gap between input concepts and output concepts 8 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S Output sphere data Meta servers for P r o c e s s M e t a d a ta 9 Qualtity metadata • Define quality standards (required output quality) • Rules to measure resulting quality 10 Focused on Energy Statistics • Energy statistics are normal statistics: logic stages sceme and metadata fully applicable to energy statistics • Some elements specific for energy statistics: 11 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S National energy policy International EU (en stat reg) IEA (ESM) UN (IRES) Output sphere data Meta servers for P r o c e s s M e t a d a ta 12 Conceptual metadata Output world • Classification Energy products (IRES chapter 3 (SIEC); InterEnerStat, ESM) • Classification Energy Flows, Energy balance (IRES chapter 5 and 8; InterEnerStat, ESM) • Classification economic activity (ISIC, NACE) • Joint Annual Quest, JODI, MOS etc • Units of measurement (Joule, toe, tonnes, kWh etc.) (IRES chapter 4) • Caloric values (IRES chapter 4) 13 Conceptual metadata Input world Questionnaires • Design energy questionnaires Neth: commodity/energy balance format Registrations (‘administrative data’) • Definitions of variables and type of units Neth: client files energy companies (unit= connection adress) 14 Meta servers for C o n c e p t u a l M e t a d a ta TheStatistical StatisticalProcess Process The Input-world Output-world Input Variables D e s i g n R E S P O N D E N T S R E G I S T R A T I O N S Micro Level Input-register BaseLine Output Variables S t a g e : Micro level Output-register MicroBase 4 Macro- Level Cube StatBase metadata Macro- Level Output Database StatLine 2 3 U 1 S E 5 6 Data collection & data entry Input sphere 7 Aggregation & Disclosure Control Editing & imputation 8 Throughput sphere Implementation stage: 9 Selection & Tabulation Publication & Dissemination R S National energy policy International EU (en stat reg) IEA (ESM) UN (IRES) Output sphere data Meta servers for P r o c e s s M e t a d a ta 15 Relation to IRES • Attempt to link IRES chapters with a logical place in the business arcitecture sceme: 16 Chapters IRES Meta servers for C o n c e p t u a l M e t a d a ta 3, 4, 5, 8 Statistical Process The Statistical Process 6b? The Input-world Output-world Input Variables 6a? R E S P O N D E N T S R E G I S T R A T I O N S Output Variables 6a? D e s i g n Micro Level Input-register BaseLine 6b? S t a g e : Micro level Output-register MicroBase Macro- Level Cube StatBase metadata - Level Macro Output Database StatLine U 1,2 S E 7a 11 R Data collection & data entry Aggregation & Disclosure Control Editing & imputation Input sphere Selection & Tabulation stage: S 10 Throughput sphere Implementation Publication & Dissemination Output sphere data 9 Metadata Quality 7b? Meta servers for P r o c e s s M e t a d a ta 17 summary (1) • Statistical process seen as a logical output oriented cycle • Starts with users (identification user needs) • Ends with users (provide desired statistical results) • Structured chain of successive actions • Delimited by intermediate products (logical databases) • Supported by the coherent use of metadata (conceptual, process, quality) 18 summary (2) • Important to make clear distinction between input world and output world • Explicitly bridge the gap between input concepts and output concepts: -input definitions output definitions -observation units statistical units 19 summary (3) • Logical model applicable to energy statistics (as being normal statistics) • IRES may be structured according to the lines of the model • Which seems not completely be the case right now! (in particular: distinction input/output world) 20