HLG Wiki Home

Site Map

Modernisation Committees

MC Org. Framework & Evaluation

[MC Production & Methods]

[MC Products & Sources]

[MC Standards]

HLG resources

Virtual Standards Helpdesk

Events

HLG documentation

 ...

Other Resources

 ...

The Role of Big Data in the Modernisation of Statistical Production

 

 

 

I.               Background

 

1.               The importance of the relationship of Big Data to the official statistics industry has been raised in a number of arenas during recent years.  At a High-Level Seminar on Streamlining Statistical Production and Services, held in St Petersburg, 3-5 October 2012, participants called for a strategically-focused document aimed at heads and senior managers of statistical organizations, outlining the issues, challenges and opportunities that Big Data poses for official statistics.  The resulting paper [1] discussed definitions and sources, and identified challenges in the areas of legislation, privacy, financial aspects, management, methodology and technology. 

 

2.               Subsequently the April 2013 meeting of the UNECE Expert Group on the Management of Statistical Information Systems (MSIS) once again identified Big Data as a key challenge for official statistics, and called for the High-Level Group for the Modernisation of Statistical Production and Services (HLG) to focus on the topic in its plans for future work.  A temporary task team composed of representatives of 13 national and international statistics organizations was convened to formulate the present project proposal.

 

3.               This project is important for the HLG’s broad programme of modernisation of statistical production.  As a component of the modernisation programme, it will contribute to the goals of international harmonisation and collaborative approaches to new challenges, improved efficiency of statistical production, and the modification of products and production methods to meet changing user needs.  The HLG’s strategy document [2] states that “products and services must become easier to produce, less resource-intensive, and less burdensome on data suppliers” and that “new and existing products and services should make use of the vast amounts of data becoming available, to provide better measurements of new aspects of society”.  The project is aligned with these aspirations since it focuses on new sources, new methods, new outputs, and ways to tackle the issues surrounding these.

 

4.               This project outline includes the objectives, scope and content of this project, as well as some practical project management issues.

 

II.               Project objectives

 

5.               The project has three main objectives:

 

      To identify, examine and provide guidance for statistical organizations to act upon the main strategic and methodological issues that Big Data poses for the official statistics industry

      To demonstrate the feasibility of efficient production of both novel products and ‘mainstream’ official statistics using Big Data sources, and the possibility to replicate these approaches across different national contexts

      To facilitate the sharing across organizations of knowledge, expertise, tools and methods for the production of statistics using Big Data sources.

 

III.               Scope

 

6.               This project concerns the role of Big Data in the modernisation of official statistical production. It will tackle strategic and practical issues that are multi-national in nature, rather than those that are specific to individual organizations or national sources.  It will not attempt to identify a comprehensive list of all possible sources or uses of Big Data, nor can it hope to ascertain all the issues and challenges, let alone solve them, since these are broad ranging and constantly evolving. While the project does involve a practical component and a consideration of methodological issues, its aim is not to focus on the technical details of analysis of Big Data, which are covered by other national and international projects.

 

7.               By including representatives of many national and international statistical organizations in the task team that formulated this project proposal, and by continuing to consult with these and other partners throughout, the project aims to be complementary to other initiatives and to avoid duplication of efforts. It also aims to be as relevant as possible to organization-specific needs and concerns.

 

IV.               Contents

 

8.               This project comprises the four work packages outlined below.  As a precursor to the project, the following activities are being undertaken by the temporary task team and the UNECE secretariat:

 

      Formulation of a classification scheme for Big Data sources and identification of the attributes of these sources that are relevant to their use in the production of official statistics

      Development of a repository with examples of sources being used, products being created and other activities being undertaken by statistical organizations, according to the classification and attributes identified above.  This repository will also include case studies for organizations intending to use similar sources or undertake similar projects

      Initial specification and costing of the ‘sandbox’ environment described under work package 2 below.

 

Work Package 1: Issues and Methodology

 

9.               This work package involves an analysis of the major strategic questions posed by the emergence of Big Data. It will require, first of all, more concrete definitions of the various terms. The work package will require very broad inputs from across the statistical community and hence will begin with gathering input through electronic consultation and virtual meetings.

 

10.               The work package will expand on, and seek to address the major challenges listed in the HLG paper ‘ What does Big Data mean for Official Statistics?’:

 

      Legislative: how to access and use data?

      Privacy: how to manage public trust and acceptance of data re-use and linking to other sources?

      Financial: what are the potential costs and benefits of using Big Data?

      Management: what policies are necessary for the effective management and protection of the data?

      Methodological: how to manage data quality? Are current statistical methods and models suitable for Big Data?

      Technological: what are the issues related to information technology?

 

11.               It will also address a variety of issues and questions identified by the task team, including (but not limited to) the following:

 

      How can we assess the suitability of Big Data sources for the production of official statistics?

      Can we identify best practices or guidelines for the major methodological issues relating to Big Data? E.g.:

o         Methods for reducing data volume

o         Methods for noise reduction

o         Methods for obtaining information on statistical concepts (text mining, classification methods, etc.)

o         Methods for determination of population characteristics, e.g. determining the population of users of social media services through analysis of words or phrases that are highly correlated with certain demographic characteristics

o         Assessing the applicability of models

      Should Big Data be treated as homogeneous, or do they require different treatment according to the role they play in the production of official statistics?

o         Experimental uses

o         Complementing existing statistics e.g. benchmarking and validity checking;

o         Supplementing existing sources, permitting the creation of entirely new statistics;

o         Replacing existing sources and methods

      Are there ‘quick wins’, applicable beyond Big Data, such as data storage, technology, advanced analytics, methods and models which could transform our thinking in relation to the production of official statistics more generally?

      How should statistical organizations react to the novel idea that in a Big Data world there are no ‘bad’ data (they all tell us something)?

      Do we need a research question before exploring a Big Data source, or should we just experiment and innovate to see what is possible?

      What becomes of the time series in a world where data sources and uses may become more transient?

      What is the demand for new types of statistical information, given the new possibilities?

      How should statistical organizations approach the need to ‘educate’ (or re-educate) staff and users?

 

12.               The output from this work package will take the form of recommendations, good practices and guidelines, developed through broad consultation of experts throughout the official statistics community, and coordinated by expert task teams.  The material will be collated in an electronic environment such as a wiki.  Such an environment will allow the guidelines to function as a ‘living document’, permitting timely updating as circumstances change.  The task of maintaining the content after its initial formulation will be overseen by the HLG’s Modernisation Committee on Products and Sources.

 

Work Package 2: Shared computing environment (‘sandbox’) and practical application

 

13.               This work package will form the practical element of the project.  A web-accessible environment for the storage and analysis of large-scale datasets will be created and used as a ‘sandbox’ for collaboration across participating institutions.  One or more free or low-cost, internationally-relevant datasets will be obtained and installed in this environment, with the goal of exploring the tools and methods needed for statistical production and the feasibility of producing Big Data-derived statistics and replicating outputs across countries.  Simple configurations with tools and data will, whenever possible, be released in ‘virtual machines’ that partners will be able to download in order to test them within their own technical environments.

 

14.               The sandbox will be used as a platform for proving concepts in two related strands:

 

(a)               Statistics:

      the possibility of producing valid and reliable statistics from novel sources, including the ability to produce statistics which correspond in a predictable and systematic way with existing ‘mainstream’ products, such as price statistics

      the cross-country applicability of new analytical techniques and sources, such as the analysis of data from social networking websites.  This will be done by attempting to reproduce the results of a national project in other countries

 

(b)               Tools:

      the efficiency of various software tools for large-scale processing and analysis

      the applicability of the Common Statistical Production Architecture (CSPA – under development) to the production of statistics using Big Data sources.

 

Work Package 3: Training and dissemination

 

15.               This work package will ensure that the conclusions reached in the two preceding work packages are shared broadly throughout the statistical world and beyond.  This will be done through a variety of means, including:

 

(a)               establishing and maintaining a central location for documentation and information sharing on the UNECE wikis, including detailed documentation arising from work package 2

(b)               preparation of electronic demonstrations of tools and results, for example in the form of Webex presentations and Youtube videos which can be disseminated widely.  Identification of existing electronic resources and online training materials is also included in this strand

(c)               a workshop in which the results of work package 2 will be presented to members of the various of expert groups involved in the HLG’s modernisation programme. This would be held back-to-back with the annual workshop on modernisation of statistics at which all these expert groups are represented (likely to be November 2014).

  Work Package 4: Project management and coordination

 

16.               This work package comprises the necessary project management activities to ensure the successful delivery of the other three work packages.

 

V.               Definition of success

 

17.               Overall, this project will be successful if it results in an improved understanding within the international statistical community of the opportunities and issues associated with using Big Data for the production of official statistics. Success criteria for the individual work packages are:

 

      Work package 1: a consistent international view of Big Data issues and solutions, documented and released through a public web site

      Work package 2: recommendations on appropriate tools, methods and environments for processing and analysing different types of Big Data, and a report on the feasibility of establishing a shared approach for using multi-national Big Data sources.

      Work package 3: exchange of knowledge and ideas between interested organizations and a set of standard training materials

      Work package 4: the project is completed on schedule, and delivers results that are of value to the international statistical community.

 

VI.               Expected costs

 

18.               The following table shows an estimate of the minimum resources and other costs needed to deliver the different work packages. Each organization involved in the project will be expected to cover the costs of their participation (including wages and any travel expenses for participants).

 

Work Package

Estimated resources

Source of resources

Other costs
(in US Dollars)

1: Issues and methodology

8 person months

Volunteer NSOs plus UNECE Secretariat

Possible travel costs if a workshop or sprint session is needed

2: Shared computing environment & practical applications

12 person months

Volunteer NSOs plus UNECE Secretariat

Costs associated with renting a shared space and acquiring data and tools (max $10,000?)

Possible travel costs if a workshop or sprint session is needed

3: Training & dissemination

4 person months

Volunteer NSOs plus UNECE Secretariat

Up to $1,000 for costs associated with preparing and disseminating training materials

4: Project management

6 person months

A project manager working in the UNECE Secretariat. Input from Executive Board  and HLG members (in their role as project sponsors)

Up to $500 for telecommunications and other incidentals

Travel costs for project events

Total

30 person months

UNECE Secretariat (9 person months)

NSO / International organization staff (21 person months)

Up to $11,500 for miscellaneous costs

Possible consultancy costs

Travel costs of experts

 

 

VII.               Timetable

 

19.               The project will aim to complete the activities described by the end of 2014.  There are, however, various unknowns which may affect the timetable:

 

      The availability of resources from national and international statistical organizations to support this project – if the necessary resources are not available, either the timetable will need to be extended, or the outputs will need to be re-defined (in terms of quality or quantity or both)

      The availability of project management and support resources in the UNECE Secretariat – to meet the resource requirements of this project will require the continuation of the current extra-budgetary post in the UNECE secretariat, through additional donor funding. As above, if this is not forthcoming, either the timetable will need to be extended, or the outputs will need to be re-defined.

 

20.               All four work packages will run throughout the year, though substantial work should be completed by mid-November so that outcomes can be reported and demonstrated at the HLG Workshop.

 

VIII.               Project governance

 

21.               The project sponsor is the HLG. This is the group that has ultimate responsibility for signing off the project deliverables. In practice, this responsibility will be delegated to the Executive Board.

 

22.               A project manager will have day-to-day responsibility for the running of the project, providing regular updates and signalling any issues to the Executive Board as necessary.

 


Annex: Specification for Work Package 2

 

I.                       Overall aim

 

This work package will be deemed successful if it results in strong, well-justified and internationally-applicable recommendations on appropriate tools, methods and environments for processing and analysing different types of Big Data, along with a report on the feasibility of establishing a shared approach for using Big Data sources that are multi-national or for which similar sources are available in different countries.

 

The value of the work package, in the context of the overall goals of the project on the Role of Big Data in the Modernisation of Statistical Production and in relation to the overarching strategy of the HLG, derives from its international nature.  While individual statistical organizations can experiment with the production of official statistics from Big Data sources (and many are currently doing so or have already done so), and can share their findings and methods with other organizations, this work package will be able to do the same in a more open and collaborative setting.  The work package will draw on the international nature and/or international ownership and management of many Big Data sources, and will capitalise on the collective bargaining power of the statistical community acting as one in relation to such large transnational entities.  The work package will contribute to the overall value of the project by providing a common methodology from the outset, precluding the need for post-hoc efforts to harmonise methodology in the future.

 

II.                    Specific objectives

 

Successful completion of the work package will entail evaluation of the feasibility of the following propositions; and insofar as it is found that the propositions are feasible, it will demonstrate and document in broad yet practical terms how the actions would be achievable in statistical organizations.

 

1.      'Big Data' sources can be obtained (in a stable and reliably replicable way), installed and manipulated with relative ease and efficiency on the chosen platform, within technological and financial constraints that are realistic reflections of the situation of national statistical offices

2.      The chosen sources can be processed to produce statistics (either mainstream or novel) which conform to appropriate quality criteria –both existing and new– used to assess official statistics, and which are reliable and comparable across countries

3.      The resulting statistics correspond in a systematic and predictable way with existing mainstream products, such as price statistics, household budget indicators, etc.

4.      The chosen platforms, tools, methods and datasets can be used in similar ways to produce analogous statistics in different countries

5.      The different participating countries can share tools, methods, datasets and results efficiently, operating on the principles established in the Common Statistical Production Architecture.

While the first objective is to examine these propositions (the 'proof of concept'), a second objective is to then use these findings to produce a general model for achieving the goal of producing statistics from Big Data, and to communicate this effectively to statistical organizations.  Hence, all processes, findings, lessons learned and results will be recorded and will feed into work package 3 for dissemination and training activities.  In particular, experiences and best practices for obtaining data will be detailed for the benefit of other organizations.

 

III.                  Basis for the Recommendations

 

The recommendations given in this annex were arrived at by the deliberations of a team of experts representing eight countries or international organizations, in consultation with the broader expert task team on Big Data.

The task team considered a wide range of alternative possibilities for tools, datasets and statistics and assessed them against various criteria.  These included the following:

 

Tools

      Whether or not the tools are open source

      Ease of use for statistical office staff

      Possibilities for interoperability and integration with other tools

      Ease of integration into existing statistical production architectures

      Cost

      Availability of documentation

      Availability of online tutorials and support

      Training requirements, including whether or not a vendor-specific language has to be learned

      The existence of an active and knowledgeable user community. 

 

Statistics

      At least one statistic that corresponds closely and in a predictable way with a mainstream statistic produced by most statistical organizations

      One or more short term indicators of specific variables or cross-sectional statistics which permits the study of the detailed relationships between variables

      One or more statistics that represent a new, non-traditional output (i.e. something that has not generally been measured by producers of official statistics, be it a novel social phenomenon or an existing one where the need to measure it has only recently arisen).

 

Datasets

      Ease of locating and obtaining data from providers

      Cost of obtaining data (if any)

      Stability (or expected stability) over time

      Availability of data that can be used by several countries, or data whose format is at least broadly homogeneous across countries

      The existence of ID variables which enable the merging of big data sets with traditional statistical data sources.



IV.                 Recommendations and Resource Requirements

 

The task team recommends that this work package proceed according to the items detailed in each row of the following table:

 

Aspect

Recommendations

Links to further information (where applicable)

Processing environment

HortonWorks Hadoop distribution to be installed on a cluster provided by a volunteering statistical organization.

http://hortonworks.com/

Processing tools/software

The Pentaho Business Analytics Suite Enterprise Edition will be deployed under a free trial license obtained for the purpose of the project (for an initial period of six months with the possibility of renewal up to one year).

 

Pentaho Business Analytics Suite Enterprise Edition provides a unified, interactive, and visual environment for data integration, data analysis, data mining, visualization, and other capabilities. Pentaho's Data Integration component is fully compatible with Hortonworks Hadoop and allows 'drag and drop' development of MapReduce jobs.

 

Additional tools such as R and RHadoop will be installed alongside the Pentaho Suite.

http://www.pentahobigdata.com/

http://www.r-project.org/

https://github.com/RevolutionAnalytics/RHadoop/wiki

Datasets & statistics to be produced with them (or feasibility of production to be demonstrated)

One or more from each of the categories below to be installed in the sandbox and experimented with for the creation of appropriate corresponding statistics:

      Transactional sources (from banks/telecommunications providers/retail outlets) (to enable the recreation of standard official statistics in easiest possible way, minimising as far as possible potential hindrances to access, etc.)

      Sensor data sources

      Social network sources, image or video-based sources, other less-explored sources (to enable the creation of 'new' statistics).

 

Human resource requirements

A task team will need to be identified at the outset of the project, composed of experts whose time is volunteered in kind by their respective organizations for the duration of the work package.  The project manager's first task will be to identify the number of members required, the requisite skills and the amount of time to be committed by task team members to enable the work to progress.

 

 

 



V.                    Expected timeline

 

Obtain and install necessary hardware, software etc.

January-March 2014

Assemble task team to lead sandbox work

January 2014

All those interested in participating will be encouraged to do so, but a task team will be required to steer the work, ensuring objectives are pursued and processes are documented.

 

      Set up Hortonworks system architecture

o         ensure that minimum system requirements are met (hardware recommendations, operating system requirements, browser requirements, software requirements, database requirements) 

o         decide on deployment type ( one master host and two slaves?) http://hortonworks.com/wp-content/uploads/downloads/2013/06/Hortonworks.ClusterConfigGuide.1.0.pdf ,  

o         install, configure, and deploy the cluster 

o         test the cluster

 

      Set up Pentaho suite

o         configure Pentaho for Hadoop distribution and version         

o         test the configuration

 

      Set up R with RHadoop

o         test the configuration.

 

Undergo training of task team to ensure familiarity with technical tools and start collaboration between team members

April-June 2014

      Utilisation of online documents, tutorials, demonstration videos etc.

      Potential running of a training session (conditional upon hosting and/or financial support from a participating organization), which could be undertaken alongside another Big Data event to save costs for participants.

 

Obtain requisite datasets and undertake analyses in sandbox
 

July-October 2014
 

      Obtain and install data sets (minimum of one from each category outlined in preceding section) Note: process of obtaining datasets that are not freely available (whether paid or not) should be begun at the onset of the project, in order to have them available by this stage of the work.

      For each dataset:

o         study availability of variables

o         analyse the representativeness of the statistical figures

o         study other statistical figures available

o         produce some statistics

o         document all processes and results on an ongoing basis.

 

Produce a general model for achieving the goal of producing statistics from Big Data, to communicate effectively with statistical organizations

November-December 2014

      Document findings

      Incorporate documented results into dissemination materials and activities.


 

 

 


[1] What does Big Data mean for Official Statistics? available at http://www1.unece.org/stat/platform/pages/viewpage.action?pageId=77170622

[2] Strategy to Implement the Vision of the HLG available at http://www1.unece.org/stat/platform/display/hlgbas/HLG+Strategy