Neuchâtel Terminology Model:Classification database object types and their attributes

Revision 2013 and its relation to GSIM

Prepared by Debra Mair, Tim Dunstan and Alice Born, Statistics Canada and Anne Gro Hustoft, Statistics Norway,

And the Working Group on Revision to the Neuchâtel Terminology Model

METIS 2013 Work Session (modified for EG on International Statistical Classifications) May 13, 2013

Background: The Neuchâtel terminology model: Classifications (version 2.1)

Developed by the Neuchâtel group - Denmark, Sweden, Norway, Switzerland, the Netherlands and run Software Werkstatt

Objective: to clarify some basic concepts and to arrive at a common terminology for statistical classifications

Defines the key concepts relevant to structuring classification metadata and provides the conceptual framework for the development of a classification database

Lists and describes the typical object types of a classification database, and the attributes connected with each object type

Version 2.1 published in 2004

2

Classification object types and attributes

Classification family Classification Classification version Classification variant Classification index Correspondence table Classification level Classification item Item change Case law Classification index entry

Attributes:CodeOfficial titleAlternative titlesExplanatory notesGeneral noteIncludesIncludes alsoExcludesLevel numberGeneratedCurrently validValid fromPast eventsFuture eventsChanges from previous versionUpdatesClassification version/variantParent itemSub-itemsLinked itemsCase lawsIndex entriesFootnotes

Why revise now?

Many countries have at least partially implemented the model and have expressed desire for revisions

Need for revision was discussed at 2011 METIS Workshop METIS Steering Group contacted the UN Expert Group on

International Statistical Classifications to work on the revision Result: a joint working group Generic Statistical Information Model (GSIM) drew heavily

from the Neuchâtel Terminology Model – Concepts Group revealed the need to clarify and more completely explain

certain parts of the Neuchâtel Model

4

Steps in revision process to date

Working group membership: Australia, Austria, Canada, France, New Zealand, Norway, Portugal, Sweden, the United States, Eurostat, ILO, UNECE, and UNSD

Questionnaire sent out to 25 countries and international organizations, members of the revision group plus others implementing the Neuchâtel Terminology Model

18 responses received Results showed interest in revisions, revealed key areas for revision

and provided suggestions for potential revisions The proposals received and others developed by working group

members are currently under consideration by the working group

5

Type of revisions under consideration

Definition changes Object deletions Attribute deletions, additions Clarifications, additional examples Database restructuring Mapping to GSIM

6

Classifications

Neuchâtel terminology model

Classification Family

Classification

Classification Version

Classification Level

Classification Item

Item Change

Correspondence Table

Correspondence Item

Classification Variant

Classification Index

Classification Index Entry

1 *

1

1

1

1

*

*

*

*

**

*

*

*

*

*

* *

*

*

*

*

1

1

1

1

1

11

1

1

sub_items

past_eventsfuture_events

source_items

target_items

target_version

source_version

Case law

Potential removal or “demotion” to attribute

Objects to be modified

Definition changes: classification version

Current definition of classification version:• A new version of a classification differs in essential ways from

the previous version. Essential changes are changes that alter the borders between categories, i.e. a statistical object/unit may belong to different categories in the new and old version. Border changes may be caused by creating or deleting categories, or moving part of a category to another. The addition of case law, changes in explanatory notes or in the titles do not lead to a new version.

Feedback• Clarification to the types of changes that could occur without

requiring recognition of a new version and the minimum change required for a new version.

Definition changes: classification variant

A variant is currently defined as having two characteristics:

1. At least one level of the variant is also present in its base classification version.

2. The variant and base classification version are based on similar grouping criteria.

Feedback• definition is too restrictive• fails to acknowledge the relationship to the base version of a

classification built from its classification items, but introducing a new grouping criterion (e.g., ICT variant of ISIC)

• does not recognize as variant classifications built from different levels of the base version.

Criteria considered for distinguishing between classification versions and classification variants

Registration status• Has the previous version been superseded by the introduction of

the new classification? Purpose

• Does the new classification serve the same purpose or does it serve some special purpose?

Structural changes• What is the nature of the structural changes made to the first

classification in order to create the second?

Proposed revised definition for:Classification Version

A classification version is a list of mutually exclusive categories representing the version-specific values of the classification variable. If the version is hierarchical, each level in the hierarchy is a set of mutually exclusive categories. A classification version has a certain normative status and is valid from a particular date for a period that may or may not be specified. A new version is created when a classification version is superseded by the introduction of a new classification version that differs in essential ways from the previous version. Essential changes are changes that alter the borders between categories, i.e. a statistical object/unit may belong to different categories in the new and old versions. Border changes may be caused by creating or deleting categories, or moving part of a category to another. These changes can occur at any level of the classification. The addition of case law, changes in explanatory notes or in the titles do not lead to a new version.

  It should be noted that if a classification is superseded by a new version, the

two versions will likely serve the same objective or purpose.

  Floating classifications are considered to constitute one version.

Proposed revised definition for:Classification variant

A variant is built from the classification items of a base classification version. These classification items need not all be found at the same level as in the base classification version. To these classification items, one or more new levels may be added. This can include extending the base classification version with one or several new levels at the bottom of the base classification version, creating a new lowest level.  

variants are typically developed to serve a specific purpose Removed from definition: variant and base classification version

are based on similar grouping criteria

Database re-structuring - example

Classification Item(attribute: index entries)

Classification Index(attribute: index entries)

Classification Index Entry(list of all entries)

• Index entry appears three times in the database structure - classification item, classification index and classification entry item

• proposal: index entries be dropped from classification item and classification index• Classification item and classification index be added as attributes to classification index entry• identify classification versions and variants to which the index entry is associated

Common Generic lndustrialised Statistics

GSBPM GSIM

Methods Technology

Business Concepts Information Concepts

Statistical HowTo Production HowTo

conc

eptu

alpr

actic

al

Common Generic lndustrialised Statistics

GSBPM GSIM

Methods Technology

Business Concepts Information Concepts

Statistical HowTo Production HowTo

conc

eptu

alpr

actic

al

Activity and Design

Data Collection

Processes

Unit Data Structures

Dimensional Data Structures

Dissemination

Classifications

Variables

Population and Units

Specify needs

Generic Statistical Business Process Model (GSBPM)

Other relevant standards

Geospatial standards

DDI SDMX

GSIMConceptual model

Implementationstandards

Neuchâtel TerminologyModel

Mapping Neuchâtel and GSIM

17

10 out of 17 objects map 1 to 1 between Neuchâtel and GSIM- Neuchâtel has more detail- Some conceptual differences

Classifications

Neuchâtel terminology model – object graph

Classification Family

Classification

Inheritence1 * 1:N relationship

Classification Version

Classification Level

Classification Item

Item Change

Correspondence Table

Correspondence Item

Classification Variant

Classification Index

Classification Index Entry

1 *

1

1

1

1

*

*

*

*

**

*

*

*

*

*

* *

*

*

*

*

1

1

1

1

1

11

1

1

sub_items

past_eventsfuture_events

source_items

target_items

target_version

source_version

Generic Statistical Information Model (GSIM) – object map

ClassificationClassification

Family

Classification Scheme

Classification Version

Classification Variant

Level

Correspondence Table

MapClassification

Item

groups

grou

ps

hashas

has

contains

groups

grou

ps

contains

maps

maps

Next steps

Reach agreement on set of proposed changes by the WG, METIS and EG on classifications

Working group finalizes the revisions by the end of the year, and determine the appropriate way to publish the revised model

Address remaining issues on mapping to GSIM (and other standards) and stewardship of the model

Points for discussion

A key question for this project is the potential relationship between Neuchâtel mode and GSIM• Should this project aim to produce a new version of the

Neuchâtel Model for Classifications or might an expanded GSIM take in the information objects of Neuchâtel?

Based on the decisions from point one, is there a need to address the issue of governance and stewardship of the Neuchâtel model. Currently, there is no formal owner of the model, and this is the first attempt to revise it. – At the 2013 METIS Work Session, it was agreed that a working sub-

group under the GSIM Implementation WG be created to look at incorporating the objects from Neuchâtel Terminology into GSIM, and that membership comprise both metadata and classification experts