How to Select the Right
Dimensions of Data
Quality
Includes 60 dimensions of data quality
and their standardized definitions
1
Colophon
Authors:
Andrew Black (Van Nederpelt & Black)
Peter van Nederpelt (Van Nederpelt & Black)
Reviewers:
Fred Dijk (Scamander)
Aris Prins (Premium)
Version: See Version History
Printed of 15-11-2020 14:41
© DAMA NL Foundation. All rights reserved.
2
Table of contents
Tables 4
Figures 4
1. Introduction 5
1.1 The importance of data quality 5
1.2 Purpose of this document 5
1.3 Audience of this document 5
1.4 How did the document come about? 5
1.5 Release policy 5
1.6 Reading guideline 5
2. Definitions and examples 6
2.1 What is data quality? 6
2.2 What is a data concept? 6
2.3 What is a dimension? 7
2.4 Combination of a dimension and a data concept 8
2.5 What is a requirement? 8
3. How to select the right dimensions of data quality? 9
3.1 Step 1: Determine which dimensions of data quality are important for the data
under consideration 9
3.2 Step 2: Determine whether a dimension contributes sufficiently to a higher
objective 9
3.3 Step 3: Prioritize the dimensions 9
3.4 Step 4: Establish indicators and associated measurement methods for the
selected dimensions 9
Appendix 1: Dimensions of data quality 10
Appendix 2: Combinations of dimensions and data categories 17
Appendix 3: Elaborated dimensions of data quality 20
Accuracy 21
Availability 21
Clarity 22
Completeness (1) 22
Completeness (2) 22
Consistency 23
Currency 23
Punctuality 24
Timeliness 24
Traceability 25
4
Tables
Table 1: Examples of combinations of dimensions and data concepts ................................. 8
Table 2: Examples of requirements of data quality ...................................................................... 8
Table 3: Definitions of dimensions of data quality ...................................................................... 11
Table 4: Combinations of dimensions and data categories...................................................... 17
Table 5: Definitions of concepts .......................................................................................................... 27
Figures
Figure 1: Data concepts in a data model ............................................................................................ 7
Figure 2: Relationship between data concepts and dimensions ........................................... 30
Figure 3: Artists impression of the real world and data world .............................................. 31
5
1. Introduction
1.1 The importance of data quality
Data plays an increasingly important useful role in our society. Dependence on data for
many activities and processes is increasing. Quality of data is therefore of growing
importance and should be managed.
Bad quality data puts an organisation at risk. It can lead to bad decisions, unsatisfied
customers, unsatisfied data consumers, fines due to non-compliance, hidden costs
(rework), bad reputation, unsatisfied employees, and lack of interoperability.
1.2 Purpose of this document
The purpose of this document is to present an approach to selecting the dimensions of
data quality that best apply to a specific situation. This is the first step to control or
improve data quality.
It offers, therefore, a list of 60 dimensions of data quality and corresponding,
standardized definitions (Appendix 1). The relationship with existing definitions can
be found in our exhaustive research report Dimensions of Data Quality (Black & Van
Nederpelt, 2020).
1.3 Audience of this document
The report is meant for everyone who is involved in management of data quality,
particularly those preparing to apply data quality dimensions in practice.
1.4 How did the document come about?
This document is an initiative of the Data Quality working group of DAMA-NL. This
working group drew up a research paper about dimensions of data quality (Black, Van
Nederpelt, 2020). Subsequently, the present report has been derived from this paper.
Finally, it was submitted to the DAMA community for comment and published.
1.5 Release policy
The first version of the report was published in September 2020. New versions will be
compiled as needed. Proposals for changes can be made via info@dama-nl.org or the
authors info@vannederpeltblack.nl.
1.6 Reading guideline
Chapter 1 describes the purpose and use of this document.
Chapter 2 explains some key concepts.
Chapter 3 presents the steps to select the right dimensions of data quality.
Appendix 1 shows all 60 dimensions of data quality and their definitions.
Appendix 2 presents logical combinations of dimensions of data quality and data
categories.
Appendix 3 elaborates twelve common dimensions of data quality.
Appendix 4 defines concepts used in the report.
Appendix 5 shows diagrams of the data concept system.
Appendix 6 contains references.
6
2. Definitions and examples
This chapter presents some important concepts and their definitions.
2.1 What is data quality?
ISO 9000:2015 defines quality as:
Quality is the degree to which inherent characteristics of an object meet
requirements.
We derive from this definition the following definition of data quality:
Data quality is the degree to which dimensions of data meet requirements.
Note 1: The term characteristics in the definition of ISO 9000 is substituted by
dimensions, because this term is more common in data management.
Note 2: The adjective inherent is left out, because also extrinsic dimensions are
relevant in data management such as availability.
Note 2: Data take various forms: data concepts. Each dimension of data quality is
defined in relation to such a data concept.
2.2 What is a data concept?
A data concept is defined as:
Data concept is a form by which data is structured and organised in an
information system.
Some examples of data concepts are dataset, data file, record, attribute, and data
value.
Data concepts and their definitions can be found in Appendix 4.
A subset of data concepts is shown in a data model in Figure 1.
7
Name
Nationality
Height in cm
Johnson
GB
180
Figure 1: Data concepts in a data model
In Appendix 5, two other diagrams of data concepts are presented.
2.3 What is a dimension?
Dimension is a measurable feature of a data concept.
This definition is derived from the definition from ISO 9000. This standard defines a
characteristic as a feature of an object.
The term dimension is used to make the connection to dimensions in the
measurement of physical objects (e.g., length, width, height). Examples of dimensions
are accuracy, completeness, and timeliness.
The term dimension in this context should not be confused with its use in the context
of business intelligence where it refers to a category for summarizing or viewing data.
Data file
Attribute
Data Value
Entity Type
Record
Data Item
8
2.4 Combination of a dimension and a data concept
A dimension and a data concept should be a logical combination. Examples of these
combinations are presented in Table 1. An illogical combination is, e.g., accuracy of a
data file.
Table 1: Examples of combinations of dimensions and data concepts
Dimension
Data concept
Accuracy
Data values
Completeness
Records
Completeness
Data values
Referential integrity
Data files (tables)
Definitions of dimensions of data quality in this report are composed of a combination
of a dimension and a data concept. In everyday language we tend to only mention the
dimension without the associated data concept. See also Figure 2 in Appendix 5.
2.5 What is a requirement?
Requirement is a need or expectation that is stated, generally implied or
obligatory (ISO 9000).
Examples of requirements are shown in Table 2. In the context of data quality,
requirements can be made specific by target values of indicators that are associated
with dimensions of data quality.
Table 2: Examples of requirements of data quality
Dimension
Data
concept
Requirement
Accuracy
Data values
The names in a customer file should be more than
96% correctly spelled.
Completeness
Records
The product file should contain 99,5% of the
products that the company sells.
Referential
integrity
Data files
All employees in the employee file should be
linked to a department file.
It should be noted that requirements for dimensions of data quality are context
dependant and should by established by their stakeholders. You cannot state in
general that the quality in all cases should be as high as possible because unnecessary
costs may be incurred.
9
3. How to select the right dimensions of data quality?
This chapter describes how dimensions of data quality can be selected. These actions
are the first steps in a procedure to control or improve data quality. These steps are:
• Determine which dimensions of data quality are important for the data under
consideration.
• Determine whether a dimension contributes sufficiently to a higher objective.
• Prioritize the selected dimensions
• Establish indicators and associated measurements methods for the selected
dimensions
These steps are elaborated below.
The persons who have interests in the data should be involved in the selection
process. Namely those who are responsible, accountable, consulted, and informed
(RACI) about data quality according to their roles.
3.1 Step 1: Determine which dimensions of data quality are important for the data
under consideration
Determine the category of the data. Examples of data categories are master data,
reference data, transactional data, basic registers, and statistical output.
Determine which dimensions are important for the data category. Table 4 in
Appendix 2 indicates which dimensions are candidates for a specific data category.
3.2 Step 2: Determine whether a dimension contributes sufficiently to a higher
objective
Determine whether a dimension contributes sufficiently to a higher objective, i.e. to
some business goal. The contribution must be large enough to make it worthwhile to
select the dimension.
The following are examples of such objectives:
A. Satisfaction of customers and other stakeholders
B. Quality of the product or service delivered by the organisation
C. Public confidence in the organisation
D. Reputation of the organisation
E. Interoperability between organisations
F. The level of data quality management costs relative to the costs of emergency data
quality repairs and the risks of fines due to non-compliance
G. Efficiency of the processes of all partners in a data processing chain
H. Compliance of the organisation with laws, regulations, and requirements of
regulators.
I. Data driven decisions making
See Appendix 1 for 60 dimensions of data quality and their definitions.
3.3 Step 3: Prioritize the dimensions
Rank the dimensions in order of priority. Put the dimension with the best cost-benefit
ratio on the first place, thereby keeping control over your costs.
3.4 Step 4: Establish indicators and associated measurement methods for the
selected dimensions
Establish indicators for the selected dimensions. Appendix 3 shows possible
indicators for some common dimensions.
Establish a measurement method for each indicator.
10
Appendix 1: Dimensions of data quality
This Appendix defines sixty dimensions of data quality. These dimensions can be
found in various sources related to data management. Table 3 shows the dimensions,
the associated data concept, and their definitions in alphabetical order.
The last column presents the unit of measure:
Unit of measure
Remark
%
Percentage
Number
Absolute number
Grade
Only the perception of people about the
dimension can be measured. A grade
can be number on a scale of 0-10 or 1-5.
Boolean
Yes/no or true/false
Duration
Expressed in seconds, minutes, hours,
days, weeks or months.
Story
The value of the dimension cannot be
expressed in a number and should be
explained in a ‘story’.
The following principles have been applied in compiling the definitions of the
dimensions of data quality:
▪ The list has been made as complete as possible.
▪ Definitions that already exist have been used as much as possible.
▪ The definitions meet the requirements of ISO 704. This standard is about defining
terms in general. For example, a definition should not be too long and should not
contain examples.
▪ The definition always starts with 'the degree to which...'.
▪ A dimension is always part of something. We call it a data concept (e.g. attribute,
record, or data file).
▪ The data concepts together form a data concept system. These data concepts are
also defined and visualised. See Appendix 4 and 5.
▪ Dimensions of data quality can be classified by data concept.
11
Table 3: Definitions of dimensions of data quality
Nr
Dimension
Data Concept
Definition
Unit of
measure
1.
Access security
Datasets
The degree to which
access to datasets is
restricted.
Grade
2.
Accessibility
Data
The ease with which
data can be consulted
or retrieved,
Grade
3.
Accuracy
Data values
The degree of
closeness of data
values to real values.
%
4.
Appropriateness
Format
The degree to which
the format is suitable
for use.
%
5.
Availability
Data
The degree to which
data can be consulted
or retrieved by data
consumers or a
process.
Grade
6.
Ability to
represent null
values
Format
The degree to which a
format allows null
values in an attribute.
Yes/No
7.
Clarity
Metadata
The ease with which
data consumers can
understand the
metadata.
Grade
8.
Coherence
Composition of
datasets
The degree to which
datasets can be
combined.
Story
9.
Comparability of
populations
Data values
The degree to which
data values
representing two
populations have the
same definition and
are measured in the
same way.
Grade
10.
Comparability
over time
Data values
The degree to which
data values over time
have the same
definition and are
measured in the same
way.
Grade
11.
Completeness
Attributes
The degree to which
all required attributes
in the dataset are
present.
%
12.
Records
The degree to which
all required records in
the dataset are
present.
%
12
Nr
Dimension
Data Concept
Definition
Unit of
measure
13.
Data files
The degree to which
all required data files
are present.
%,
Number
14.
Data values
The degree to which
all required data
values are present.
%
15.
Data values of an
attribute
The degree to which
all required data
values of an attribute
are present.
%
16.
Metadata
The degree to which
the metadata are fully
described.
%
17.
Compliance with
laws,
regulations, or
standards
Data
The degree to which
data is in accordance
with laws, regulations,
or standards.
Story
18.
Composition of
datasets
The degree to which
the composition of
datasets is in
accordance with laws,
regulations, or
standards.
Story
19.
Confidentiality
Data
The degree to which
disclosure of data
should be restricted to
authorized data
consumers.
Grade
20.
Consistency
Data values
The degree to which
data values of two sets
of attributes
▪ within a record,
▪ within a data file,
▪ between data files,
▪ within a record at
different points in
time
comply with a rule.
%
21.
Data values of a set
of attributes of a
dataset at different
points in time
(temporal
consistency)
The degree to which
the data values of a set
of attributes of a
dataset at different
points in time comply
with a rule.
%
22.
Data values of two
sets of attributes
between datasets
(across datasets)
The degree to which
data values of two sets
of attributes between
datasets comply with a
rule.
%
13
Nr
Dimension
Data Concept
Definition
Unit of
measure
23.
Data values of two
sets of attributes
between records
(cross record)
The degree to which
data values of two sets
of attributes between
records comply with a
rule.
%
24.
Data values of two
sets of attributes
within a record
(record level)
The degree to which
data values of two sets
of attributes within a
record comply with a
rule.
%
25.
Credibility
Data values
The degree to which
data values are
regarded as true and
believable by data
consumers.
Grade
26.
Currency
Data values
The degree to which
data values are up to
date.
%
27.
Equivalence
Attributes
The degree to which
attributes stored in
multiple datasets are
conceptually equal.
%
28.
Granularity
Attributes
The degree to which a
single characteristic is
subdivided in
attributes.
Story
29.
Records
The degree to which
objects are aggregated
to records.
Story
30.
Integrity
Data values
The degree of absence
of data value loss or
corruption.
%
31.
Interpretability
Data
The degree to which
data are in an
appropriate language
and units of measure.
%
32.
Latency
Data
The period of time
between the point
when the data is
created and the point
when it is available for
use.
Duration
33.
Linkability
Data files
The degree to which
records of one data file
can be correctly
coupled with records
of another data file.
%
14
Nr
Dimension
Data Concept
Definition
Unit of
measure
34.
Metadata
compliance
Data values
The degree to which
the data values are in
accordance with their
definition, format
specification and value
domain.
%
35.
Naturalness
Composition of
datasets
The degree to which
the composition of
datasets is aligned
with the real-world
objects that they
represent.
Grade
36.
Objectivity
Data values
The degree to which
the data values are
created in an unbiased
manner.
Grade
37.
Obtainability
Data
The degree to which
the data can be
acquired.
Grade
38.
Plausibility
Data values
The degree to which
data values match
knowledge of the real
world.
Story
39.
Portability
Data
The degree to which
data can be installed,
replaced, or moved
from one system to
another while
preserving the existing
quality.
Story
40.
Portability
Format
The degree to which a
format can be applied
in a wide range of
situations.
Story
41.
Precision (1)
Data values
The degree of accuracy
with which data values
are recorded or
classified.
Depends
on data or
metadata
42.
Precision (2)
Data values
The degree to which
the error in data
values spreads around
zero (in statistics).
%
43.
Punctuality
Dataset
availability
The degree to which
the period between
the actual and target
point of time of
availability of a dataset
is appropriate.
Duration
44.
Reasonability
Data pattern
The degree to which a
data pattern meets
expectations.
Grade
15
Nr
Dimension
Data Concept
Definition
Unit of
measure
45.
Recoverability
Datasets
The degree to which
datasets are preserved
in the event of
incident.
Story
46.
Redundancy
Data
The degree to which
logically identical data
are stored more than
once.
Number
47.
Referential
integrity
Data files
The degree to which
data values of the
primary key of one
data file and data
values of the foreign
key of another data file
are equal.
%
48.
Relevance
Composition of
datasets
The degree to which
the composition of
datasets meets the
needs of the data
consumer.
Story
49.
Reliability
Initial data value
The closeness of the
initial data value to the
subsequent data value.
%
50.
Reproducibility
Dataset
The degree to which a
dataset can be
recreated with the
same data values.
Story
51.
Reputation
Data
The degree to which
data are trusted or
highly regarded in
terms of their source
or content.
Grade
52.
Retention period
Datasets
The period that
datasets are available
until they can or must
be deleted.
Duration
53.
Timeliness
Dataset
availability
The degree to which
the period between
the time of creation of
the real value and the
time that the dataset is
available is
appropriate.
Duration
54.
Traceability
Data
The degree to which
data lineage is
available.
Story
55.
Uniqueness
Objects
The degree to which
objects (of the real
world) occur only once
as a record in a data
file.
%
16
Nr
Dimension
Data Concept
Definition
Unit of
measure
56.
Records
The degree to which
records occur only
once in a data file.
%
57.
Validity
Data values
The degree to which
data values comply
with rules.
%
58.
Value
Data
The degree to which
data provide
advantages from their
use.
Grade
59.
Variety
Data
The degree to which
data are available from
different data sources.
Story
60.
Volatility
Data values
The degree to which
data values change
over time.
%
Source: Black, A., Nederpelt, P. van. (2020). Dimensions of Data Quality Dimensions.
Research paper. DAMA-NL.
17
Appendix 2: Combinations of dimensions and data categories
Table 4 indicates which dimensions are candidates for selection in case of a specific
data category.
The dimensions in the column statistical output are numbered because in the
statistical domain these dimensions are usually presented in this sequence.
Column A contains the selection of DAMA-UK as expressed in the Six Primary
Dimensions for Data Quality Assessment.
Column B contains the selection as expressed in the List of Conformed Dimensions of
Data Quality.
Table 4: Combinations of dimensions and data categories
A
B
Data Category
Nr
Dimension
Data Concept
DAMA UK
CDDQ
Master Data
Reference Data
Transactional Data
Registers
Statistical Output
1
Access security
Datasets
X
X
X
X
2
Accessibility
Data
X
9
3
Accuracy
Data values
X
X
X
X
X
X
2
4
Appropriateness
Format
5
Availability
Data
6
Ability to
represent null
values
Format
7
Clarity
Metadata
X
1
X
X
X
X
10
8
Coherence
Composition of datasets
6
9
Comparability of
populations
Data values
8
10
Comparability
over time
Data values
7
14
Completeness
Data values
X
X
X
X
X
X
13
Completeness
Data files
X
15
Completeness
Data values of an attribute
X
12
Completeness
Records
X
X
X
X
X
11
Completeness
Attributes
X
16
Completeness
Metadata
X
X
X
X
1
In CDDQ clarity is called representation.
18
A
B
Data Category
Nr
Dimension
Data Concept
DAMA UK
CDDQ
Master Data
Reference Data
Transactional Data
Registers
Statistical Output
17
Compliance
with laws,
regulations, or
standards
Data
X
18
Compliance
with laws,
regulations, or
standards
Composition of datasets
X
19
Confidentiality
Data
X
20
Consistency
Data values
X
X
21
Consistency
Data values of a set of attributes of
a dataset at different points in time
(temporal consistency)
X
X
22
Consistency
Data values of two sets of
attributes between datasets (across
datasets)
X
X
23
Consistency
Data values of two sets of
attributes between records (cross
record)
X
X
23
Consistency
Data values of two sets of
attributes within a record (record
level)
X
X
25
Credibility
Data values
26
Currency
Data values
X
X
X
X
X
27
Equivalence
Attributes
29
Granularity
Records
28
Granularity
Attributes
30
Integrity
Data values
X
31
Interpretability
Data
32
Latency
Data
33
Linkability
Data files
34
Metadata
compliance
Data values
35
Naturalness
Composition of datasets
36
Objectivity
Data values
37
Obtainability
Data
38
Plausibility
Data values
40
Portability
Format
39
Portability
Data
41
Precision (1)
Data values
X
19
A
B
Data Category
Nr
Dimension
Data Concept
DAMA UK
CDDQ
Master Data
Reference Data
Transactional Data
Registers
Statistical Output
42
Precision (2)
Data values
43
Punctuality
Dataset availability
5
44
Reasonability
Data pattern
45
Recoverability
Datasets
X
X
X
X
46
Redundancy
Data
47
Referential
integrity
Data files
48
Relevance
Composition of datasets
1
49
Reliability
Initial data value
3
50
Reproducibility
Dataset
51
Reputation
Data
52
Retention
period
Datasets
X
53
Timeliness
Dataset availability
X
X
4
54
Traceability
Data
X
2
X
56
Uniqueness
Records
X
X
X
X
55
Uniqueness
Objects
X
X
57
Validity
Data values
X
X
X
X
58
Value
Data
59
Variety
Data
60
Volatility
Data values
2
In CDDQ traceability is called data lineage.
20
Appendix 3: Elaborated dimensions of data quality
In this Appendix twelve common dimensions of data quality are elaborated.
1. Accuracy
2. Availability
3. Clarity
4. Completeness of records
5. Completeness of data values
6. Consistency
7. Currency
8. Punctuality
9. Timeliness
10. Traceability
11. Uniqueness
12. Validity
For each dimension, the following items are described:
▪ Title. Name of the dimension.
▪ Long title. Name of the dimension and the associated data concept.
▪ Synonyms
▪ Related. Dimensions that are dependent on or contributes to the dimension.
▪ Definition
▪ Indicators. Possible indicators.
▪ Examples. Descriptions of non-compliance with required data quality.
▪ Notes
21
Accuracy
Title
Accuracy
Long title
Accuracy of data values
Synonym
Correctness of data values
Related
-
Definition
The degree of closeness of data values to real values.
Indicators
Percentage or number of inaccurate data values.
Examples of non-
compliance
▪ A house is located at number 120 but registered as number
12.
▪ A person is called Janssen but registered as Jansen.
▪ A farm has 7,321 chickens. It is registered as 7,321 while the
unit of measurement is thousand. It should be registered as 7.
▪ A product is located at A23 but according to the database its
location is P76.
▪ The number of unemployed people is estimated at 234.000. If
the sample is not fully representative, there will be bias or
systematic error. The size of the sample determines the
variance or random error of the estimate.
Notes
▪ The data producer or consumer must define when he/she
considers a data value as inaccurate and define criteria for
inaccuracy.
▪ The impact of an inaccuracy is different for each attribute.
▪ Generally, accuracy will be measured for individual
attributes, e.g., the accuracy of the product name.
Availability
Title
Availability
Long title
Availability of data
Synonyms
-
Related
Obtainability of data
Definition
The degree to which data can be consulted or retrieved by
data consumers or a process.
Indicators
▪ Yes or No
▪ The effort it takes to make data available (hours)
Examples of non-
compliance
▪ Data are not available because they are not processed
yet such as the number of casualties of a recent
incident.
▪ Personal data are not available to the public.
▪ Data are not available for reasons of competition.
▪ Data are not available because they are confidential or
secret.
▪ Data are not available because they not archived in a
professional manner.
Notes
Data can be partly available.
22
Clarity
Title
Clarity
Long title
Clarity of metadata
Synonyms
-
Related
Unambiguity, readability
Definition
The ease with which data consumers can understand the metadata.
Indicators
A grade (1-10)
Examples of non-
compliance
▪ The name of a file is 765897xyp.asc. This name has little
meaning.
▪ Data attribute ‘profit’ has no definition. It is not clear if it is net
or gross profit.
Notes
Other quality dimensions of metadata are completeness,
correctness, and availability.
Completeness (1)
Title
Completeness
Long title
Completeness of records
Synonyms
Coverage
Related
-
Definition
The degree to which all required records in the dataset are present.
Indicators
Percentage or number of the required records that are present.
Examples of non-
compliance
▪ Not all products are present in a product file.
▪ Not all inhabitants of a city are registered.
▪ A file of trees also contains shrubs (superfluous records)
Notes
Incomplete records are also called missing units.
Completeness (2)
Title
Completeness
Long title
Completeness of data values
Synonyms
-
Related
-
Definition
The degree to which all required data values are present.
Indicators
Percentage of the possible data values that are present.
Examples of non-
compliance
▪ In a product file the attribute supplier is not completed in every
record.
▪ In a questionnaire a respondent did not answer all questions.
Notes
Incomplete data values are also called missing values.
23
Consistency
Title
Consistency
Long title
Consistency of data values
Synonyms
-
Related
Plausibility of data values
Definition
The degree to which data values of two sets of attributes
▪ within a record,
▪ within a data file,
▪ between data files,
▪ within a record at different points in time
comply with a rule.
Indicators
Percentage of inconsistencies.
Examples of non-
compliance
▪ A company is registered in the city of Paris in the country of
Belgium.
▪ Overlap are gaps in file with address history of a person. For
example: Address A from 1 Jan 2003 – 1 May 2019 and Address
B from 1 March – until now.
Notes
-
Currency
Title
Currency
Long title
Currency of data values
Synonyms
-
Related
Timeliness of availability of data
Definition
The degree to which data values are up to date.
Indicators
Percentage of data that are up to date in a point of time.
Examples of non-
compliance
Outdated prices in the product file.
Notes
-
24
Punctuality
Title
Punctuality
Long title
Punctuality of the availability of a dataset
Synonyms
-
Related
Timeliness of the availability of a dataset
Definition
The degree to which the period between the actual and target point
in time of availability of a dataset is appropriate.
Indicators
▪ The period between the actual and target point in time of
availability of a dataset (days, hours, minutes).
▪ Percentage of times that datasets were available too late (or too
early).
Examples of non-
compliance
▪ The dataset should be available on 1 July 2020 but is released on
3 July 2020. Too late.
▪ The dataset should be available on 1 July 2020 at 10:00 am but
is released at 9.45 am. Too early.
Notes
▪ A dataset can also consist of one transaction.
▪ If no target time is agreed or planned, punctuality cannot be
measured.
Timeliness
Title
Timeliness
Long title
Timeliness of the availability of a dataset
Synonyms
-
Related
Punctuality of the availability of a dataset
Definition
The degree to which the period between the time of creation of the
real value and the time that the dataset is available is appropriate.
Indicators
Percentage of times a dataset was not available in a timely manner.
Examples of non-
compliance
▪ The date of birth of a person is available in a dataset after 23
days. It should be available within one week.
▪ Data about quarterly returns of VAT are available 3 months after
the end of the quarter. The requirement is 1 month after the end
of the quarter.
Notes
▪ Timeliness can only be measured if there is a norm for
timeliness, e.g., one week after the event.
▪ Timeliness is dependent on the duration of a process.
▪ Data can be available punctually but not timely and the other
way around.
25
Traceability
Title
Traceability
Long title
Traceability of data
Synonyms
-
Related
-
Definition
The degree to which data lineage is available.
Indicators
A grade (1-10)
Examples of non-
compliance
The source of the data is unknown.
Notes
Data lineage is metadata that identifies the sources of data and the
transformations through which it has passed up to the point of
consumption.
Uniqueness
Title
Uniqueness
Long title
Uniqueness of records
Synonyms
-
Related
Uniqueness of objects
Definition
The degree to which records occur only once in a data file.
Indicators
Percentage of duplicates in a data file.
Examples of non-
compliance
Product A occurs twice in a file.
Notes
▪ A record that occurs twice in a data file is called a duplicate.
▪ Uniqueness of object is the degree to which objects (of the real
world) occur only once as a record in a dataset.
Three different problems can occur:
a. One record with one key value occurs more than once in a
dataset (duplicate with identical key values). The two records
are not unique.
Key
Name
22
John
22
John
b. One record with more than one key value occurs more than once
in a dataset (duplicate with different key values). Object John is
not unique in the dataset.
Key
Name
22
John
37
John
c. One record has the same key as another record, and both occur
in a dataset (false duplicate). Key 22 is not unique.
Key
Name
22
John
22
Peter
26
Validity
Title
Validity
Long title
Validity of data values
Synonyms
-
Related
Accuracy of data values
Completeness of data values
Consistency of data values
Definition
The degree to which data values comply with rules.
Indicators
Percentage of data values that do not comply with rules.
Examples of non-
compliance
▪ A city that does not exist in a list of cities.
▪ A birth data that is out of range of valid birth dates.
Notes
▪ A data value can be valid but not accurate.
▪ A data value can be valid but incomplete. Absence of certain data
values may be permitted.
▪ A valid data value is part of a value domain.
▪ Consistency is about comparing two or more data values.
27
Appendix 4: Definitions of concepts and data concepts
In this Appendix concepts are defined that are relevant in this report. See Table 5.
A distinction is made between data concepts in the real world (purple) and the data
world (yellow). Other concepts (white) are more general.
Each word that appears in bold in the definition of a concept is a concept defined
elsewhere in Table 5. This way the coherence between the concepts is made visible.
Table 5: Definitions of concepts
Concept
Definition
Source
Relationships with
other concepts
Attribute
A characteristic of an entity type
about which the organisation wishes
to hold information.
-
Distinguishes entity
type
Is specified by its
name, definition,
classification and
format.
Characteristic
Distinguishing feature
ISO 9000
-
Composition of
a dataset
The way in which a dataset is made
up.
-
Concept
Unit of knowledge created by a unique
combination of characteristics
ISO 1087
-
Concept system
A set of concepts structured according
to the relations among them.
ISO 704
-
Data
A representation of facts, concepts, or
instructions in a formalized manner,
suitable for communication,
interpretation, or processing by
humans or by automatic means. (ISO
2382-4).
In: ISO
11179
-
Data category
A classification of data according to
the purpose for which it is used.
-
-
Data concept
A form by which data is structured
and organised in an information
system.
-
Has associated
dimensions
Data file
Data stored on a computer as
one unit with one name.
Cambridge
2020
Is part of a dataset.
Data item
One occurrence of an attribute
-
Contains data value
Data lineage
Metadata that identifies the sources
of data and the transformations
through which is has passed up to the
point of consumption.
-
-
Data pattern
A series of data that repeats in a
recognizable way.
Investo-
pedia
-
28
Concept
Definition
Source
Relationships with
other concepts
Data quality
Data quality is the degree to which
dimensions of data meet
requirements
Adapted
from ISO
9001
-
Data value
The value of a data item.
-
Is contained in data
item
Forms part of
record
Is within value
domain
Represents a
property of an
object
Dataset
Any organized collection of data.
Early 2011
Is composed of data
files
Dataset
availability
The degree to which a dataset can be
consulted or retrieved by data
consumers or processes.
-
Is a characteristic of
a dataset.
Dataset
composition
The way in which a dataset is made
up.
-
-
Definition
Representation of a concept by an
expression that describes it and
differentiates it from related concepts
ISO 1087
-
Dimension
Measurable characteristic.
DAMA 2017
Is associated with a
data concept.
Entity type
A thing of significance about which the
organisation wishes to hold
information
Hay 2013
Is distinguished by
attributes
Describes object
Initial data
value
A provisional data value that will be
updated by a more accurate value.
-
Is a specification of a
data value.
Format
A combination of datatype, unit of
measure and character set.
-
Is part of the
specification of an
attribute.
Metadata
Data that defines and describes other
data.
ISO 11179
-
Master Data
Data held by an organization which
describe object types that it needs to
reference in order to perform its
transactions.
-
Is an instance of
data category.
Object
Anything perceivable or conceivable.
ISO 9000
Is described by
entity type
Is characterised by
properties
Is represented by
records
29
Concept
Definition
Source
Relationships with
other concepts
Property
A feature of an object.
ISO 1087
Characterises object
Is recorded by data
value
Actually, has real
value
Register
A dataset designated by the
government in which vital data about
citizens, residents, companies,
institutions, vehicles, topography,
buildings, and addresses can be
centrally maintained.
-
Is an instance of
data category.
Statistical
output
Output from a statistical process.
-
Is an instance of
data category.
Transactional
data
Data that describes an event that takes
place as an organization conducts its
business.
-
Is an instance of
data category.
Real value
The real-life value of a property of an
object.
-
Expresses an
instance of a
property.
Reference data
Data used to categorize other data.
-
Is an instance of
data category.
Record
A logically related set of data values
that represent a (real-world) object
-
Forms part of data
file
Is composed of data
values
Value domain
A set of permissible values of an
attribute.
-
Includes data value
Source: Black, A., Nederpelt, P. van. (2020). Data concept system for Data Quality
Dimensions. Research paper. DAMA-NL.
30
Appendix 5: Diagrams
Figure 2 shows that a dimension is associated with a data concept. The definition of a
dimension of data quality is formed by the combination of a dimension and a data
concept. In the diagram, only the common dimensions are presented.
Figure 2: Relationship between data concepts and dimensions
31
Figure 3 is an artitst impression of the real world and data world.
Figure 3: Artists impression of the real world and data world
32
Appendix 6: Sources
Sources of definitions of quality dimension
Brackett, M. H. (2012). Data Resource Integration: Understanding and Resolving a
Disparate Data Resource (1
st
ed.). Bradley Beach, NJ: Technics Publications, LLC.
CDDQ. (n.d.). List of Conformed Dimensions of Data Quality. Retrieved from
https://dimensionsofdataquality.com/alldimensions
Daas, P.J.H., Ossen, S.J.L., & Tennekes, M. (2010). Determination of
Administrative Data Quality: Recent results and new developments.
Retrieved from Q2010 website:
https://q2010.stat.fi/media//presentations/special-session-
34/daas_ossen_tennekes_q2010_paper_session34_piet_daas_paper.pdf
DAMA (2017). DAMA-DMBOK. Data Management Body of Knowledge. 2
nd
Edition.
Technics Publications LLC. August 2017.
DAMA-UK (2013). The six primary dimensions for data quality assessment. October
2013.
Earley, S. (2011). The DAMA Dictionary of Data Management (2
nd
ed.). NJ: Technics
Publications LLC.
English, L. P. (1999). Improving Data Warehouse and Business Information Quality:
Methods for Reducing Costs and Increasing Profits. Hoboken, NJ: Wiley.
Eurostat. (2015). ESS Handbook for Quality Reports. (2015). Brussels, Belgium:
Eurostat.
Everest. (2010).
Fisher, Craig; Lauria, Eitel; Chengalur-Smith, Shobha; Wang, Richard (2011).
Introduction to Information Quality. Bloomington: AuthorHouse.
ISO 25012. (n.d.). Retrieved from https://iso25000.com/index.php/en/iso-25000-
standards/iso-25012
Jayawardene, J., Sadiq, S., & Indulska, M. (2015). An analysis of data quality
dimensions. Computer Science. Retrieved from
https://pdfs.semanticscholar.org/3d9a/c49f03b3e4bebae7c0e7eb20d7fde7222a9
c.pdf
Nederpelt, P.W.M. van (2009). Checklist Quality of Statistical Output. Den
Haag/Heerlen: Centraal Bureau voor de Statistiek.
Redman, T. C. (1996). Data Quality for the Information Age. Artech House on Demand.
Wang, R.Y. and Strong, D. (1996). Beyond Accuracy: What data quality means to Data
Consumers. Journal of Management Information Systems, 1996. 12(4): p. 5 – 34.
Other sources
Black, A., Nederpelt, P. van. (2020). Data concept system for Data Quality Dimensions.
Research paper. DAMA-NL.
Black, A., Nederpelt, P. van. (2020). Dimensions of Data Quality Dimensions. Research
paper. DAMA-NL.
Brackett, M. H. (2012). Data Resource Integration: Understanding and Resolving a
Disparate Data Resource (1
st
ed.). Bradley Beach, NJ: Technics Publications, LLC.
Business Directory. (2020, 1 mei). Geraadpleegd op 2 mei 2020, van
http://www.businessdictionary.com/definition/conceptual-framework.html
Cambridge English dictionary. (2020, August 1). Meanings & definitions. Cambridge
Dictionary | English Dictionary, Translations &
Thesaurus. https://dictionary.cambridge.org/dictionary/english
DAMA (2017). DAMA-DMBOK. Data Management Body of Knowledge. 2
nd
Edition.
Technics Publications Llc. August 2017.
Earley, S. (2011). The DAMA Dictionary of Data Management (2
nd
ed.). NJ: Technics
Publications Llc.
33
Hay, D. (2013). Data model patterns: Conventions of thought. Addison-Wesley.
Humbley, J., Budin, G., & Laurén, C. (2018). Languages for Special Purposes. Berlin,
Germany: De Gruyter.
Investopedia. (2015, January 7). Patterns vs. trends: What’s the
difference? https://www.investopedia.com/ask/answers/010715/what-are-
differences-between-patterns-and-trends.asp
ISO 1087 (2019). Terminology work and terminology science — Vocabulary. Vernier,
Switzerland: ISO.
ISO 11179-1 (1999). Information technology — Specification and standardization of
data elements — Part 1: Framework for the specification and standardization of
data elements. ISO
ISO 21961 (2003). Space data and information transfer systems – Data entity
dictionary specification language (DEDSL) – Abstract syntax. Retrieved from
https://www.iso.org/obp/ui#iso:std:iso:21961:ed-1:v1:en.
ISO 704 (2009). Technical Committee ISO/TC 37; Terminology and other language
and content resources. Subcommittee SC 1; Principles and Methods.
(2009). Terminology Work – Principles and Methods. Vernier, Switzerland: ISO.
ISO 9000:2015 (2015). Quality management systems – Fundamentals and vocabulary.
Delft: NNI.
ISO 9001:2015 (2015). Quality management systems – Requirements. Delft: NNI.
Jonker R. (2020, May 6). Terminologie. Retrieved from
https://labyrinth.rienkjonker.nl/lexicon/terminologie
Lexico Dictionaries (2020, August 1). Definitions, meanings, synonyms, and grammar
by Oxford dictionary on Lexico.com. Lexico Dictionaries |
English. https://www.lexico.com
Merriam-Webster. (2020, May 7). Dictionary by Merriam-Webster: America’s most-
trusted online dictionary. Retrieved from https://www.merriam-
webster.com/dictionary
Nederpelt, P.W.M. van (2012). Object-oriented Quality and Risk Management. A
practical method for quality and risk management. New York/Alphen den Rijn:
Lulu/MicroData.
Oxford. (2020, May 7). Oxford learner’s dictionaries | Find definitions, translations,
and grammar explanations at Oxford learner’s dictionaries. Retrieved from
https://www.oxfordlearnersdictionaries.com
Regoniel, P. (2015, January 5). Conceptual framework: A step-by-step guide on how to
make one. Retrieved from https://simplyeducate.me/2015/01/05/conceptual-
framework-guide/
UN/Edifact glossary. (n.d.). Retrieved from
https://www.unece.org/trade/untdid/texts/d300_s.htm
34
Version history
Version
Date
Description of the modification
Author
1.0.p1
14 August
2020
First draft
Peter
1.0.p2
20 August
Amendments and comments
Andrew
1.0.p3
20 August
Amendments and comments
processed
Peter
1.0.p4
27 August
2020
Comments Fred Dijk processed.
Peter
1.0.p5
28 Augustus
2020
Diagram edited.
Andrew
1.0.p6
28 August 20
Amendments and comments
Andrew
1.0
3 Sept 20
Comments processed
Peter
1.1.p1
19 Sept 20
Added:
• Risks of insufficient data quality
• Reference to research report
DDQ with existing definitions
• Unit of measure
• Roles and responsibilities (RACI)
• Excel spreadsheet
• Names of reviewers
• Prioritization of dimensions
Removed:
• Procedure to improve data
quality.
Peter
1.1.p1
20 Sept 20
Amendments and comments
Andrew
1.1.p3
21 Sept 20
Amendments and comments
processed. Chapter 2 reformulated.
Peter
1.1.p4
1 Oct 20
Amendments to definitions
Andrew
1.1
14 Nov 20
Amendments processed.
Peter
Active distribution per version
Version
Distribution
1.0.p1-2
Dropbox
1.0.p3
Dropbox, Fred Dijk (review)
1.0.p4
Dropbox, Reviewers.
1.0.p5-p6
Dropbox
1.0
Dropbox. Website DAMA-NL
1.1.p1
Dropbox. Fred Dijk.
1.1.p2
Dropbox
1.1.p3
Dropbox, Fred Dijk
1.1.p4
Dropbox
1.1
Dropbox. Website DAMA-NL
