Statistical Classification management with DDI-L

Kaia Kulla

Statistics Estonia

​

This work is licensed under Creative Commons Attribution 4.0 International License.

Outline

• Background
• Neuchâtel Model -> GSIM –> DDI-L
• Object types in Classification Model and its properties in DDI-L (consistent with GSIM), examples
• Classification Family and Series
• Statistical Classification
• Classification Item, Level
• Classification Index and Index Entry
• Correspondence Table, and Map
• Summary: Advantage of DDI-L

​

​

​

2

Background: Classification in DDI

• Neuchâtel Terminology Model: Classification Database object types and their attributes v 2.1 (2004; revised 2013)
• GSIM Statistical Classification Model v1.1 (2013)
• GSIM - Generic Statistical Information Model
• The first internationally endorsed reference framework for statistical information
• DDI 3.2 Code List and Copenhagen Mapping based on GSIM, (2014)
• DDI 3.3 Statistical Classification and related objects (2020)
• Classification management based on GSIM / Neuchâtel Model

​

3

GSIM Classification Model

4

3 _Object graph - 3 _Object graph - UNECE Statswiki

Classification Family

• A Classification Family is a group of Classification Series related from a particular point of view.
• The Classification Family is related by being based on a common Concept (e.g. economic activity).

​

5

Classification Family in DDI-L, example

6

Classification Series

•  A Classification Series is an ensemble of one or more Statistical Classifications, based on the same concept, and related to each other as versions or updates.
• Typically, these Statistical Classifications have the same name.

7

Classification Series in DDI-L, example

8

Statistical Classification

• The Statistical Classification is exhaustive and structured set of mutually exclusive and well-described categories, often presented in a hierarchy that is reflected by the numeric or alphabetical codes assigned to them, used to standardize concepts and compile statistical data.

[Eurostat’s Concepts and Definitions Database]

​

• A Statistical Classification is a set of categories which may be assigned to one or more variables registered in statistical surveys or administrative files, and used in the data collection, production and dissemination of statistics.

9

Types of Statistical Classification

• A Statistical Classification may be
• versionable
• valid from a particular date for a period that may or may not be specified
• floating
• a validity period should be defined for all Classification Items which will allow the display of the item structure and content at different points of time

​

​

10

Structure of the Classification

• A Statistical Classification may have
• a flat, linear structure or
• may be hierarchically structured, such that �all categories at lower Levels are �sub-categories of categories at �the next Level up.

​

• Categories in Statistical Classifications �are represented in the information �model as Classification Items.

​

11

Statistical Classification in DDI-L, example

12

Classification Item

• A Classification Item represents a Category at a certain Level within a Statistical Classification.
• It defines the content and the borders of the category.
• An object/unit can be classified to one and only one Classification Item at each Level of a Statistical Classification.

13

Classification Item in DDI-L, example

14

Level

• A Statistical Classification has a structure which is composed of one or several Levels.
• A Level often is associated with a Concept, which defines it.
• In a hierarchical classification the Classification Items of each Level but the highest are aggregated to the nearest higher Level.
• A linear classification has only one Level.

15

Level in DDI-L, example

16

Classification Index

• A Classification Index is an ordered list (alphabetical, in code order etc.) of Classification Index Entries.
• A Classification Index can relate to one particular or to several Statistical Classifications.
• A Classification Index shows the relationship between text found in statistical data sources (responses to survey questionnaires, administrative records) and one or more Statistical Classifications. 
• A Classification Index may be used to assign the codes for Classification Items to observations in statistical collections. 

17

Classification Index in DDI-L, example

18

Classification Index Entry

• A Classification Index Entry is a word or a short text (e.g. the name of a locality, an economic activity or an occupational title) describing a type of object/unit or object property to which a Classification Item applies, together with the code of the corresponding Classification Item.
• Each Classification Index Entry typically refers to one item of the Statistical Classification.
• Although a Classification Index Entry may be associated with a Classification Item at any Level of a Statistical Classification, Classification Index Entries are normally associated with items at the lowest Level.

19

Classification Index Entry in DDI-L, example

20

Correspondence Table

• A Correspondence Table expresses the relationship between two Statistical Classifications.
• These are typically:
• two versions from the same Classification Series; 
• Statistical Classifications from different Classification Series;
• a variant and the version on which it is based; or,
• different versions of a variant.
• In the first and last examples, the Correspondence Table facilitates comparability over time.
• Correspondence relationships are shown in both directions.

21

Correspondence Table in DDI-L, example

22

Map

• A Map is an expression of the relation between a Classification Item in a source Statistical Classification and a corresponding Classification Item in the target Statistical Classification.
• The Map should specify whether the relationship between the two Classification Items is partial or complete.
• Depending on the relationship type of the Correspondence Table, there may be several Maps for a single source or target item.

23

Mapping in DDI-L, example

24

Summary

25

Advantages of DDI-L

​

​

​

​

26

Time saving

Quality control

Reusability

Consistency

Life-cycle tracking

Provenance

References to the material used

• GSIM Statistical Classifications Model
• 4 _Object types and attributes - 4 _Object types and attributes - UNECE Statswiki

​

• DDI 3.3 Classification / Item types
• Classification — DDI Data Documentation 3.3 0.1 documentation (ddialliance.github.io)

​

27