Nilgün DORSAN Turkish Statistical Institute

TURKISH STATISTICAL INSTITUTE

Metadata and Standards Department

International Seminar on Modernizing Official Statistics: Meeting Productivity and New Data Challenges, 24-25

October 2013

Nilgün DORSANTurkish Statistical

Institute

How the standards based modernization of statistical

production serves dissemination?


Metadata and Standards Department 22

What makes a What makes a statistical statistical organisation efficient? organisation efficient? 11• Use of administrative records• Optimization of data collection• Standardization of business processes• Management of human resources• Decisions at each stage of statistical

processing cycle1 From the presentation made by Lidia Bratanova, UNECE



Turkish Turkish Statistical Statistical InstituteInstitute Main data provider in Turkish Statistical

System Coordination role in Official Statistical

Programme (63 organization, 287 subjects) Central and 26 regional offices, 1672

permanent and 1920 temporary: total 3592 staff

The vision for 2012-2016 is “to establish a user focused and sustainable statistical system based on international standards”

Principles of official statistics and Principles of official statistics and Turkish Statistical LawTurkish Statistical Law

Trust in statistics

Trust in statistical organizations



Why standardization is important?Why standardization is important?To define the production line To establish comparable, harmonized and

integrated systemsTo portray the statistical studiesTo define which outputs of the surveys are inputs

of which surveysTo consolidate the concepts within the institutionTransfer of institutional culture to future

generationsEffective organizational structureBase for the statistical metadata systemBenchmark or peer review processes with

other organizations within the statististical system



Some of the objectives and Some of the objectives and strategies on standardizationstrategies on standardizationTurkStat 2012-2016 Strategic Plan: •Development of strategic management approach and management of the knowledge •Creation and development of process management•Establishing metadata system•Use of international standards (NACE, ISCED, PRODCOM, etc.)



2 Design

3Build

4Collect

5 Process

6Anaylse

7 Dissemination

1 Specify needs

3.4Finalise production

system

3.3 Test production system

3.2 Integrate production

System with othersystems

3.1Build and enhance

production system components

Determine business plan

1.3Establish output

objectives

1.2Consult and confirm

need

1.1Determine need for information

2.5Design production

system and work flows

2.4Design process and

analysis methodology

2.3Design data collection

methodology

2.2Design frame, register

and sample methodology

Design statistcalProducts and

outputs

4.4 Finalise collection

4.3 Run collection

4.2Set up collection

4.1Establish frame

and registers, selectsample

5.1 Classify and code

5.2 Micro-edit

5.4Imputation

5.3 Macro-control

6.6 Finalisecontent

6.1Evaluate

the information forİts effect

6.2 Produce statistics

6.3Quality assure

statistics

6.4Examine and

evaluate statistics

6.5 Prepare statistics

fordissemination

7.4Manage user

demands

7.3 Manage publishingfor dissemination

product

7.2Produce

dissemination product

7.1Update

disseminationsystems

1.4Check data availability

Data Collection

Thematic Units

Methodology and Information Systems

TurkStat Draft Model for GSBPM (Level 1 and 2)Statistics Production Process

Dissemination

5.5Calculate weights

and derive variables



Actions for Actions for sstreamlining the treamlining the sstatistical tatistical pproductionroduction• Process modeling and standardization studies launched in 2010

• TurkStat Draft Statistical Businesss Process Model in 2011

• Bilateral negotiations with the technical units• Processes for 273 products and services• Standardization plan for metadata and processes in 2011Under 7 main processes,33 processes in 2nd level,

106 sub processes in 3rd level and thousands of activities under these levels.



Reference and Structural Reference and Structural Metadata Metadata • Metada template prepared in DDI format• Training of technical units• Code lists used in the surveys were

standardized across all business surveys and all household surveys, wherever possible.

125 projects in DDI format 103 standard code lists 143 codes in global repository



Where dissemination begins? Where dissemination begins? Main statistical processes and systems

Input

Raw data

Raw data

Microdata

Microdata

Statistical data(Macro data)

Statistical data

Output

Collect Process Analyse Disseminate

ProductionDatabase

ProductionDatabase

Institutional Database

Dissemination Database



OutputsOutputs• News bulletins• Dissemination databases• Statistical tables• Anonymized micro data• Printed / electronic publications

Disseminated through web page, teletext, micro data research center, paper, CD, Social media



• In 2012, 267 news bulletin in 80 subjects• Dissemination databases in 58 subjects

including detailed time series data• 1 336 statistical tables• Metadata

12

Main dissemination channel Main dissemination channel : www.tuik.gov.tr: www.tuik.gov.tr



Challenge facing statistical producersChallenge facing statistical producers



Challenges of the dataChallenges of the data• New data needs in statistical domains• Abundance of data sources and data flows

from different registers• Management of the data and data sources• Non standardised codes• Data integration• Micro data exchange and transperancy• Data confidentiality• Effective use of the data and its quality



Without a good description of the needs and effective

design of the processes, you can not get quality outputs.



2 Design

3Build

4Collect

5 Process

6Anaylse

7 Dissemination

1 Specify needs

3.4Finalise production

system

3.3 Test production system

3.2 Integrate production

System with othersystems

3.1Build and enhance

production system components

Determine business plan

1.3Establish output

objectives

1.2Consult and confirm

need

1.1Determine need for information

2.5Design production

system and work flows

2.4Design process and

analysis methodology

2.3Design data collection

methodology

2.2Design frame, register

and sample methodology

Design statistcalProducts and

outputs

4.4 Finalise collection

4.3 Run collection

4.2Set up collection

4.1Establish frame

and registers, selectsample

5.1 Classify and code

5.2 Micro-edit

5.4Imputation

5.3 Macro-control

6.6 Finalisecontent

6.1Evaluate

the information forİts effect

6.2 Produce statistics

6.3Quality assure

statistics

6.4Examine and

evaluate statistics

6.5 Prepare statistics

fordissemination

7.4Manage user

demands

7.3 Manage publishingfor dissemination

product

7.2Produce

dissemination product

7.1Update

disseminationsystems

1.4Check data availability

Data Collection

Thematic Units

Methodology and Information Systems

TurkStat Draft Model for GSBPM (Level 1 and 2)Statistics Production Process

Dissemination

5.5Calculate weights

and derive variables



Specify needs

For an efficient data production and meeting the user needs some questions arise…

• What users wants?• What is our sources?• Do we need extra data source?• Which data source best meets the needs?• How we manage the big data?



• Data integration from multiple sources• Data structure• Management of metadata and meta

information• Statistical outputs• Variables• Concepts and definitions• Classifications, code lists• Questionnaire design

• Design of architecture

Design



Structural Structural MetaMetadata data ComponentsComponents

Metadata Editor

Productiondatabase

InstitutionalDatabase

Dissemination

databases



Necessary for an efficient production system.Without this process you can not manage and integrate the data.

Build



• Production of statistics / indexes• Evaluation of data quality• Seasonal adjustments• Evaluation of response rates• Benchmarks for other periods and data sources• Application of data confidentiality rules• Decision of the dissemination of data level • Metadata preparation for outputs• Preparation of the data for the dissemination

Analyse



• Feedback of the users and update of the specifying needs

• Update dissemination system for relevant, timely and accesible data

• Development of the dissemination channels

Disseminate



TurkStat’s main principles TurkStat’s main principles for disseminationfor dissemination• Maximum data:open and free access

• Transperancy, clarity• Timeliness and punctuality • Comparability



Data Data ConfidentialityConfidentialityConfidentiality rules determined in Law

No.5429 (Articles 4, 12, 13 and 14) and related regulation

Data Confidentiality Board Decisions on conflictsConfidentiality Commitment Document is

signed by every staff


Metadata and Standards Department 25

Production process: Continuous improvement, Production process: Continuous improvement, monitoring, evaluation (quality management)monitoring, evaluation (quality management)

Missing

coverag

e

Respondent

serrors……..

Plan

DoCheck

ActI have to rotate this wheel

Incomplete

administrativ

e records

Data incompatible with

classifications and

standards

International Comparability

Up to date and trusted data



Future Future pplans for lans for streamliningstreamlining• Finalize the statistical business process

model integrated with a strong statistical metadata system.

• Provide a strong infrastructure for quality management system.

• A leading model for other producers of official statistics.

• Develop metadata infrastructure



Thank you for your Thank you for your attention. attention.

Any question?Any question?

Nilgün DORSAN Turkish Statistical Institute

Documents