=Paper=
{{Paper
|id=Vol-1184/paper9
|storemode=property
|title=daQ, an Ontology for Dataset Quality Information
|pdfUrl=https://ceur-ws.org/Vol-1184/ldow2014_paper_09.pdf
|volume=Vol-1184
|dblpUrl=https://dblp.org/rec/conf/www/Debattista0A14
}}
==daQ, an Ontology for Dataset Quality Information==
https://ceur-ws.org/Vol-1184/ldow2014_paper_09.pdf
daQ, an Ontology for Dataset Quality Information
Jeremy Debattista Christoph Lange Sören Auer
University of Bonn / University of Bonn / University of Bonn /
Fraunhofer IAIS, Germany Fraunhofer IAIS, Germany Fraunhofer IAIS, Germany
name.surname@iais- math.semantic.web auer@cs.uni-bonn.de
extern.fraunhofer.de @gmail.com
ABSTRACT dataset’s quality. This statement is incorrect as the five star scheme
Data quality is commonly defined as fitness for use. The problem serves as a guide in order to lead data to reach increasing levels of
of identifying the quality of data is faced by many data consumers. interlinkage, openness and standardisation. Therefore, although it
To make the task of finding good quality datasets more efficient, we is favourable to have a five star linked open dataset, dataset quality
introduce the Dataset Quality Ontology (daQ). The daQ is a light- issues might still be unclear to data publishers. Various works pro-
weight, extensible vocabulary for attaching the results of quality mote quality measurements on linked open data [5, 8, 13]. Zaveri
benchmarking of a linked open dataset to that dataset. We dis- et al. [15] goes a step further by providing a systematic literature
cuss the design considerations, give examples for extending daQ review.
by custom quality metrics, and present use cases such as browsing To put the reader into the context of this work, we introduce a
datasets by quality. We also discuss how tools can use the daQ to use case:
enable consumers find the right dataset for use.
Bob is a medical doctor and a computer enthusiast.
During his free time he is currently working on a mo-
Categories and Subject Descriptors bile application that would help colleagues to find out
The Web of Data [Vocabularies, taxonomies and schemas for the possible medicines to treat patients. Currently he is
web of data] experimenting with a popular data management plat-
form, hoping to find a suitable medical dataset for
General Terms reuse. Fascinated by the views, especially the faceted
filtering techniques available on this platform, Bob is
Documentation, Measurement, Quality, Ontology particularly interested in reputable medical datasets.
As he downloaded some datasets and viewed them in a
1. INTRODUCTION visualisation tool, he found out that most of the data is
The Linked (Open) Data principles of using HTTP URIs to rep- either irrelevant for his work or contains many incor-
resent things have facilitated the publication of interlinked data on rect and inconsistent facts.
the Web, and their sharing between different sources. The com-
monly used Resource Description Framework (RDF) provides both Data quality is commonly defined as fitness for use [14]. The
publishers and consumers of Linked Data with a standardised way problem of identifying the quality of data is faced by many data
of representing data. A substantial amount of facts has already consumers. A simple approach would be to rate the “fitness” of a
been published as RDF Linked Open Data.1 These facts have been dataset under consideration by computing a set of defined quality
extracted from heterogeneous sources, which also include semi- metrics. On big datasets, this computation is time consuming, even
structured data, unstructured data, documents in markup languages more so when multiple datasets are to be filtered or compared to
such as XML, and relational databases. The use of such a variety of each other by quality. Apart from this, the identification of the
sources could lead to problems such as inconsistencies and incom- quality of a dataset cannot be reused; other data consumers would
plete information. A common open data user’s perception2 is that have to do this process all over again.
the five star scheme for open data3 would automatically approve a To make the task of finding good quality datasets more efficient,
1
we introduce the Dataset Quality Ontology (daQ). The daQ is a
http://lod-cloud.net light-weight ontology that allows datasets to be “stamped” with its
2
This is following a discussion with some Open Data enthusiasts quality measures. In contrast to related vocabularies that represent
at the University of Malta.
3 quality requirements (cf. Section 5), our ontology allows for ex-
http://5stardata.info
pressing concrete, tangible values that represent the quality of the
data. Having this metadata available in the datasets enables data
publishers and consumers to automatically perform tedious tasks
such as filtering and comparing dataset quality. With the Dataset
Quality Ontology we aim to add another star to the LOD five star
scheme, for data that is not just linked and open, but of a high qual-
ity.
1.1 Terminology
Copyright is held by the author/owner(s).
LDOW2014, April 8, 2014, Seoul, Korea. To prepare the reader for the discussions carried out within the
following sections, we define some terminology, paraphrasing def- 2.1 Cataloguing and Archiving of Datasets
initions by Zaveri et al. [15]: Software such as CKAN4 , which is best known for driving the
datahub.io platform5 , makes datasets accessible to consumers by
• A Quality Dimension is a characteristic of a dataset relevant providing a variety of publishing and management tools and search
to the consumer (e.g. Availability of a dataset). facilities. A data publisher should be able to upload to such plat-
forms, whilst on the other hand the platform should be able to au-
• A Quality Metric is a procedure for measuring a data quality
tomatically compute metadata regarding the dataset’s quality. With
dimension, which is abstract, by observing a concrete quality
the knowledge from this metadata, the publisher can improve the
indicator. This assessment procedure returns a score, which
quality of the dataset. On the other hand, having quality metadata
we also call the value of the metric. There are usually multi-
available for candidate datasets, consumers would be given the op-
ple metrics per dimension; e.g., availability can be indicated
portunity to discover certain quality aspects of a potential dataset.
by the accessibility of a SPARQL endpoint, or of an RDF
dump. The value of a metric can be numeric (e.g., for the 2.2 Dataset Retrieval
metric “human-readable labelling of classes, properties and
Tools for data consumers, such as CKAN, usually provide fea-
entities”, the percentage of entities having an rdfs:label or
tures such as faceted browsing and sorting, in order to allow
rdfs:comment) or boolean (e.g. whether or not a SPARQL
prospective dataset users (such as Bob, introduced in the previous
endpoint is accessible).
section) to search within the large dataset archive. Using faceted
browsing, datasets could be filtered according to tags or values of
• A Quality Category is a group of quality dimensions in
metadata properties. The datasets could also be ranked or sorted
which a common type of information is used as quality in-
according to values of properties such as relevance, size or the
dicator (e.g. Accessibility, which comprises not only avail-
date of last modification. Figure 1 shows a mockup of a modi-
ability but also dimensions such as security or performance).
fied datahub.io user interface to illustrate how quality attributes and
Grouping the dimensions into categories helps to arrange a
metrics could be used in a faceted search with ranking.
clearer breakdown of all quality aspects, given their large
With many datasets available, filtering or ranking by quality can
number. Zaveri et al. have identified 23 quality dimensions
become a challenge. Talking about “quality” as a whole might
(with almost 100 metrics) and grouped them into 6 cate-
not make sense, as different aspects of quality matter for differ-
gories [15].
ent applications. It does, however, make sense to restrict quality-
Whenever it is necessary to subsume all of these three concepts, we based filtering or ranking to those quality categories and/or dimen-
will use the term Quality Protocols. sions that are relevant in the given situation, or to assign custom
weights to different dimensions, and compute the overall quality
1.2 Structure of this Paper as a weighted sum. The daQ vocabulary provides flexible filter-
The remainder of this paper is structured as follows: in Section 2 ing and ranking possibilities in that it facilitates access to dataset
we discuss use cases for the daQ vocabulary. Then, in Section 3 quality metrics in these different dimensions and thus facilitates the
and 4 we discuss the vocabulary design and give examples of how (re)computation of custom aggregated metrics derived from base
this vocabulary can be extended and used. Finally, in Section 5 we metrics. To keep quality metrics information easily accessible, we
give an overview of similar ontology approaches before giving our strongly recommend that each dataset contains the relevant daQ
final remarks in Section 6. metadata graph in the dataset itself.
Alexander et al. [1] provide the readers with a motivational use
case with regard to how the voID ontology (cf. Section 5) can help
2. USE CASES with effective data selection. The authors describe that a consumer
Linked Open Data quality has different stakeholders in a myr- can find the appropriate dataset by basing a criteria for content
iad of domains, however the stakeholders can be cast under either (what is the dataset mainly about), interlinking (to which other
publishers or consumers. dataset is the one in question interlinked), and vocabularies (what
Publishers are mainly interested in publishing data that others vocabularies are used in the dataset). The daQ vocabulary could
can reuse. The five star scheme, which we propose to extend by a give an extra edge to “appropriateness” by providing the consumer
sixth star for quality, defines a set of widely accepted criteria that with added quality criteria on the candidate datasets.
serve as a baseline for assessing data reusability. The reusability
criteria defined by the five star scheme and by quality metrics are 3. VOCABULARY DESIGN
largely measurable in an objective way. Thanks to such objective The Dataset Quality Ontology (daQ) is a vocabulary for attach-
criteria, one can assess the reusability of any given dataset without ing the results of quality benchmarking of a linked open dataset to
the major effort of, for example, running a custom survey to find that dataset. The idea behind daQ is to provide a core vocabulary,
out whether its intended target audience finds it reusable. (Such a which can be easily extended with additional metrics for measuring
survey may, of course, still help to get an even better understanding the quality of a dataset. The benefit of having an extensible schema
of quality issues.) is that quality metrics can be added to the vocabulary without major
Without an objective rating that is easy to determine, data con- changes, as the representation of new metrics would follow those
sumers – both machine and human – may find it challenging to as- previously defined.
sess the quality of a dataset, i.e. its fitness for use. Machine agents, daQ uses the namespace prefix daq, which expands to http:
e.g. for discovering, cataloguing and archiving datasets, may lack //purl.org/eis/vocab/daq.
the computational power required to assess some of their quality di- The basic and most fundamental concept of daQ is the Quality
mensions, e.g. logical consistency. Tools for human end users, such Graph (Figure 2 – Box A), which is a subclass of rdfg:Graph. daQ
as semantic web search engines [12] or Web of Data browsers [4,
4
10, 11], do not currently focus on quality when presenting a list of http://www.ckan.org
5
search results or an individual dataset. http://www.datahub.io
Figure 1: datahub.io Mockup having Quality Attributes available in the Faceted Browsing and Ranking
A computedOn
rdfg:Graph QualityGraph rdfs:Resource
hasDimension hasMetric value
Category Dimension Metric
dateComputed requires
B xsd:dateTime rdfs:Resource
Classes Properties
rdfg:Graph Concept in an existing ontology
Subclass of
Dataset Concept in proposed ontology
requires Abstract Property
Abstract concept in proposed ontology. (Not for direct use)
Category hasDimension
Not for direct use. Property in proposed ontology
Undefined object: Concept or Literal
Figure 2: The Dataset Quality Ontology (daQ)
instances are stored as RDF Named Graph [7] in the dataset whose value’s daq:hasValue range is inherited by the actual metric’s
quality has been assessed. Named graphs are favoured due to attribute. A metric might also require additional information
(e.g. a gold standard dataset to compare with). Therefore,
• the capability of separating the aggregated metadata with re- a concrete metric representation can also define such prop-
gard to computed quality metrics of a dataset from the dataset erties using subproperties of the daq:requires abstract prop-
itself; erty. Each metric can record the date when it was actually
• their use in the Semantic Web Publishing vocabulary [6] to computed using the daq:dateComputed.
allow named graphs to be digitally signed, thus ensuring
trust in the computed metrics and defined named graph 4. USING THE ONTOLOGY
instance. Therefore, in principle each daq:QualityGraph We start this section by first showing how the daQ vocabulary
can have the following triple :myQualityGraph can be extended, and then proceed by giving general recommen-
swp:assertedBy :myWarrant . dations on how to publish daQ metadata records with datasets. We
then continue by showing how a typical daQ instance is represented
The daQ ontology distinguishes between three layers of abstrac-
and we give some SPARQL examples to demonstrate how a data
tion, based on the survey work by Zaveri et al. [15]. As shown
consumer (including tools such as the filtering UI presented in Sec-
in Figure 2 Box B, a quality graph comprises of a number of dif-
tion 2.2) can query the daQ vocabulary and a graph instance. We
ferent Categories (C), which in turn possess a number of quality
conclude this section by describing an application which will use
Dimensions (D). A quality dimension groups one or more com-
the proposed vocabulary to filter and rank datasets.
puted quality Metrics (M ). To formalise this, let G represent the
named Quality Graph (daq:QualityGraph), C = {c1 , c2 , . . . , cx } 4.1 Extending daQ
is the set of all possible quality categories (daq:Category), D =
The classes of the core daQ vocabulary can be extended by more
{d1 , d2 , . . . , dy } is the set of all possible quality dimensions
specific and custom quality metrics. In order to use the daQ one
(daq:Dimension) and M = {m1 , m2 , . . . , mz } is the set of all
should define the quality metrics which characterise the ”fitness for
possible quality metrics (daq:Metric); where x, z, y ∈ N, then:
use“ [14] in a particular domain. However, we are currently in the
D EFINITION 1. process in defining the quality dimensions and metrics described
in [15], as the standard set of quality protocols for Linked Open
G ⊆ C, Data in daQ.7 Extending the daQ vocabulary means adding new
C ⊂ D, quality protocols that inheriting the abstract concepts (Category-
Dimension-Metric). In Figure 4 we show an illustrative example
D ⊂ M;
of extending the daQ ontology (TBox) with a more specific quality
Figure 3 shows this formalisation in a pictorial manner using attribute, i.e. the RDF Availability Metric as defined in [15], and
Venn diagrams. an illustrative instance (ABox) of how it would be represented in a
Quality metrics can, in principle, be calculated on a collection of dataset.
statements - datasets or graphs. This vocabulary allows a data pub- The Accessibility concept is defined as an rdfs:subClassOf
lisher to create multiple graphs of quality metrics for different data. the abstract daq:Category. This category has five dimensions,
For example, if one dataset consists of a number of graphs, qual- one of which is the Availability dimension. This is defined
ity metrics can be defined for each graph separately. The property as a rdfs:subClassOf daq:Dimension. Similarly, RDFAvailabil-
daq:computedOn with domain daq:QualityGraph allows a data ityMetric is defined as a rdfs:subClassOf daq:Metric. The
publisher to define a quality graph for different rdfs:Resources. daq:hasValue property is also extended with a sub-property called
The resource should be the URI of a dataset (including instances daq:hasDoubleValue which types its range as xsd:double. The spe-
of void:Dataset6 ) or an RDF named graph. cific properties hasAvailabilityDimension and hasRDFAccessibili-
tyMetric (sub-properties of daq:hasDimension and daq:hasMetric
3.1 Abstract Classes and Properties respectively) are also defined (Figure 4). The advantage of extend-
This ontology framework (Figure 2) has three abstract class- ing the abstract ontology concepts in Figure 2 Box B is that the
es/concepts (daq:Category, daq:Dimension, daq:Metric) and domain and range of daq:hasDimension and daq:hasMetric are re-
four abstract properties (daq:hasDimension, daq:hasMetric, stricted to the appropriate quality protocols.
daq:hasValue, daq:requires) which should not be used directly in a Extensions by custom quality metrics do not need to be made in
quality instance. Instead these should be inherited as parent classes the daQ namespace itself; in fact, in accordance with LOD best
and properties for more specific quality protocols. The abstract practices, we recommend extenders to make them in their own
concepts (and their related properties) are described as follows, as- namespaces. Extending the daQ vocabulary with additional met-
suming the definitions given in Section 1.1: rics assumes that their exact semantics (such as how they are to
be computed) is understood by some software implementation, be-
daq:Category represents the highest level of quality assessment. cause daQ is intended to remain light-weight and thus not capable
A category groups a number of dimensions. of expressing such semantics by its own means. Therefore, a user
extending the daQ would not normally need to specify the techni-
daq:Dimension – In each dimension there is a number of metrics. cal requirements of the quality metric, although pointers to such
requirements descriptions can be given via specialisations of the
daq:Metric – The smallest unit of measuring a quality dimension
daq:requires abstract property.
is a metric. Each metric has a value, representing a score
for the assessment of a quality attribute. Since this value is 4.2 Publishing daQ Metadata Records
multi-typed (for example one metric might return true/false
7
whilst another might require a floating point number), the The metrics defined so far can be found under the namespace URI
given above, or in the source files at https://github.com/
6
http://www.w3.org/TR/void/#dataset diachron/quality/blob/master/vocab/.
G
C1 Cx
D1 D2 Dx
M1 M2 M3 Mx
Figure 3: Venn Diagram depicting Definition 1
Figure 4: Extending the daQ Ontology – TBox and ABox
query that retrieves and ranks all datasets by the Entity Trust [15]
Dataset publishers should offer a daQ description as an RDF metric. This query is useful for consumers who would require a
Named Graph in their published dataset. Since such a daQ meta- visible ranking of the datasets.
data record requires metrics to be computed, it is understandable
select ?catInst, ?dimInst where {
that it is not easy to author manually. Therefore, as suggested in ?qualGraph a daq:QualityGraph .
Section 2, publishing platforms should offer such on-demand com- graph {
putation to dataset publishers. Another possible tool would be a ?category rdfs:subClassOf daq:Category .
?property rdfs:subPropertyOf daq:hasDimension .
pluggable platform which calculates quality metrics on datasets. }
Since the daQ vocabulary can be easily extended by custom quality graph ?qualGraph {
metrics, as shown in the previous section, such a pluggable en- ?catInst a ?category ;
?property ?dimInst .
vironment would allow users to import such custom metrics. One }
must keep in mind that the computation of metrics on large datasets }
might be computationally expensive; thus, platforms computing
dataset quality must be scalable. Listing 2: A SPARQL query retrieving all Category and
Dimension instances from a daq:QualityGraph
4.3 Representing Quality Metadata Instances
Listing 1 shows an instance of the daq:QualityGraph in a select ?metricInst where {
dataset. ex:qualitygraph1 is a named daq:QualityGraph. The ?qualGraph a daq:QualityGraph .
triples show that quality metrics were computed on the whole graph {
dataset. Consumers’ queryies of the dataset for daq:QualityGraph ?metric rdfs:subClassOf daq:Metric .
}
instances will resolve to the named graph. In this named graph ?qualGraph {
graph, instances for the daq:Accessibility, daq:Availability, ?metricInst a ?metric ;
daq:EndPointAvailabilityMetric and daq:RDFAvailabilityMetric daq:doubleValue ?val .
filter(?val < 0.5)
are shown. A metric instance specifies the metric value and the }
date when it was last computed. }
# ... prefixes
Listing 3: A SPARQL query retrieving all Metrics which have
# ... dataset triples their value (double) < 0.5
ex:qualityGraph1 a daq:QualityGraph ;
daq:computedOn <> . select ?dataset where {
?qualGraph a daq:QualityGraph ;
ex:qualityGraph1 { daq:computedOn ?dataset
# ... quality triples graph ?qualGraph {
ex:accessibilityCategory a daq:Accessibility ; ?metricInst a daq:EntityTrustMetric ;
daq:hasAvailabilityDimension ex:availabilityDimension daq:doubleValue ?val .
. order by desc(?val) .
}
ex:availabilityDimension a daq:Availability ; }
daq:hasEndPointAvailabilityMetric ex:endPointMetric ;
daq:hasRDFAvailabilityMetric ex:rdfAvailMetric .
Listing 4: A SPARQL query retrieving and rank all Datasets
ex:endPointMetric a daq:EndPointAvailabilityMetric ; by the Entity Trust metric value
daq:dateComputed "2014-01-23T14:53:00"^^xsd:dateTime
;
daq:doubleValue "1.0"^^xsd:double . 4.5 The DIACHRON Project
ex:rdfAvailMetric a daq:RDFAvailabilityMetric ; The DIACHRON project (“Managing the Evolution and Preser-
daq:dateComputed "2014-01-23T14:53:01"^^xsd:dateTime vation of the Data Web”8 ) combines several of the use cases men-
;
daq:doubleValue "1.0"^^xsd:double .
tioned so far. DIACHRON’S central cataloguing and archiving hub
is intended to host datasets throughout several stages of their life-
# ... more quality triples cycle [3], mainly evolution, archiving, provenance, annotation, ci-
} tation and data quality. As a part of the DIACHRON project, we
are implementing scalable and efficient tools to assess the quality
Listing 1: A Dataset Quality Graph N3 instance of datasets. A web-based visualisation tool, to be implemented as
a CKAN plugin, will
4.4 Retrieving Metadata using SPARQL • allow data publishers to perform quality assessment on
Queries datasets, which will provide them with quality score meta-
Listings 2 and 3 show typical SPARQL queries, which could be data and also assist them with fixing quality problems;
performed by data consumers. The first query retrieves all cate-
gory and dimension instances from the quality graph. This query • allow data consumers to filter and rank datasets by multiple
could be useful, for example, for those consumers who require to quality dimensions.
visualise all categories and dimensions available in a faceted man-
The daQ vocabulary is the core ontology underlying these services.
ner. The second query retrieves all metric instances whose value
It will help these services to do their jobs, i.e. adding quality meta-
(in this case a double-precision floating point number) is less than
data to datasets, which in turn is displayed on the web frontend.
0.5. This might be useful for identifying those metrics w.r.t. which
8
the dataset needs serious improvement. Listing 4 shows a SPARQL http://diachron-fp7.eu
5. RELATED WORK of quality metrics are also being implemented, with the aim of pro-
To the best of our knowledge, the Data Quality Management viding information about the quality of big LOD datasets. This
(DQM) vocabulary [9] is the only one comparable to our approach. would allow us to create meaningful daQ Named Graph instances
Fürber et al. propose an OWL vocabulary that primarily represents at a large scale, i.e. creating quality metadata on real datasets. The
data requirements, i.e. what quality requirements or rules should be DIACHRON platform will support the daQ by ranking and filter-
defined for the data. Such rules can be defined by the user herself, ing datasets according to the quality metadata, like we sketched in
and the authors present SPARQL queries that “execute” these re- the mockup explained in Section 2.2. Having tools and platforms
quirements definitions to compute metrics values. Unlike our daQ supporting the daQ will finally allow us to test and evaluate the
model, the DQM defines a number of classes that can be used to vocabulary thoroughly, to see whether the daQ itself is of a high
represent a data quality rule. Similarly, properties for defining rules quality, i.e. fit for use.
and other generic properties such as the rule creator are specified.
The daQ model allows for integrating such DQM rule definitions 7. ACKNOWLEDGMENTS
using the daq:requires abstract property, but we consider the defi- This work is supported by the European Commission under
nition of rules out of daQ’s own scope. As discussed in Section 2, the Seventh Framework Program FP7 grant 601043 (http://
the intention of the Dataset Quality vocabulary is to enable data diachron-fp7.eu).
publishers to easily describe dataset quality so that, in turn, con-
sumers can easily find out which datasets are fit for their intended 8. REFERENCES
use. Rather than having quality rules defined using the daQ itself,
[1] K. Alexander, R. Cyganiak, M. Hausenblas, and J. Zhao.
the semantics of the custom metric concepts should be understood
Describing Linked Datasets – On the Design and Usage of
by the application implementing them. Therefore, rather than hav-
voiD, the ‘Vocabulary of Interlinked Datasets’. In WWW
ing a fixed set of classes/rules which one can extend, the daQ vo-
2009 Workshop: Linked Data on the Web (LDOW2009),
cabulary gives the freedom to the user to define and implement any
Madrid, Spain, 2009.
metrics required for a certain application domain.
[2] J. Attard, S. Scerri, I. Rivera, and S. Handschuh.
Our design approach is inspired by the digital.me Context Ontol-
ogy (DCON9 ) [2]. Attard et al. present a structured three-level rep- Ontology-based situation recognition for context-aware
resentation of context elements (Aspect-Element-Attributes). The systems. In Proceedings of the 9th International Conference
DCON ontology instances are stored as Named Graphs in a user’s on Semantic Systems, I-SEMANTICS ’13, pages 113–120,
New York, NY, USA, 2013. ACM.
Personal Information Model. The three levels are abstract concepts,
which can be extended to represent different context aspects in a [3] S. Auer, L. Bühmann, C. Dirschl, O. Erling, M. Hausenblas,
concrete ubiquitous computing situation. R. Isele, J. Lehmann, M. Martin, P. N. Mendes, B. van
The voID10 and dcat11 ontologies recommended by the W3C Nuffelen, C. Stadler, S. Tramp, and H. Williams. Managing
provide metadata vocabulary for describing datasets. The “Vocabu- the life-cycle of linked data with the LOD2 stack. In
lary of Interlinked Datasets” (voID) ontology allows the high-level Proceedings of International Semantic Web Conference
description of a dataset and its links [1]. On the other hand, the (ISWC 2012), 2012. 22
Data Catalog Vocabulary (dcat) describes datasets in data catalogs, [4] T. Berners-Lee, Y. Chen, L. Chilton, D. Connolly,
which increase discovery, allow easy interoperability between data R. Dhanaraj, J. Hollenbach, A. Lerer, and D. Sheets.
catalogs and enable digital preservation. With the daQ ontology, we Tabulator: Exploring and analyzing linked data on the
aim to extend what these two ontologies have managed for datasets semantic web. In Proceedings of the The 3rd International
in general to the specific aspect of quality: enabling the discovery Semantic Web User Interaction Workshop (SWUI06), Nov.
of a good quality (fit to use) datasets by providing the facility to 2006.
“stamp” a dataset with quality metadata. [5] C. Bizer. Quality-Driven Information Filtering in the Context
of Web-Based Information Systems. PhD thesis, FU Berlin,
Mar. 2007.
6. CONCLUDING REMARKS [6] J. J. Carroll, C. Bizer, P. J. Hayes, and P. Stickler. Semantic
In this paper we presented the Dataset Quality Ontology (daQ), web publishing using named graphs. In J. Golbeck, P. A.
an extensible vocabulary to provide quality benchmarking meta- Bonatti, W. Nejdl, D. Olmedilla, and M. Winslett, editors,
data of a linked open dataset to the dataset itself. In Section 2 we ISWC Workshop on Trust, Security, and Reputation on the
presented a number of use cases that motivated our idea, including Semantic Web, volume 127 of CEUR Workshop Proceedings.
cataloguing, archiving and filtering datasets, and that helped in de- CEUR-WS.org, 2004.
veloping the daQ ontology (Section 3). The ontology is still in its [7] J. J. Carroll, P. Hayes, C. Bizer, and P. Stickler. Named
initial phases, thus further modelling will be required in the coming graphs, provenance and trust. In A. Ellis and T. Hagino,
months to make sure that the core vocabulary covers all concepts editors, Proceedings of the 14th WWW conference, pages
required for the intended use cases. This will be possible by (i) 613–622. ACM Press, 2005.
exchanging ideas with interested LOD quality researchers, and (ii) [8] A. Flemming. Quality Characteristics of Linked Data
making sure that the vocabulary meets the standards required to be Publishing Datasources. http://sourceforge.net/
easily adapted by both data producers and consumers. apps/mediawiki/trdf/index.php?title=
We are currently in the process of giving more precise definitions Quality_Criteria_for_Linked_Data_sources,
of the quality dimensions and metrics collected in [15]. A number 2010. [Online; accessed 13-February-2014].
9
http://www.semanticdesktop.org/ontologies/ [9] C. Fürber and M. Hepp. Towards a vocabulary for data
dcon/ quality management in semantic web architectures. In
10 Proceedings of the 1st International Workshop on Linked
http://www.w3.org/TR/void/
11 Web Data Management, LWDM ’11, pages 1–8, New York,
http://www.w3.org/TR/2014/
REC-vocab-dcat-20140116/ NY, USA, 2011. ACM.
[10] A. Harth. Visinav: A system for visual search and navigation
on web data. Web Semantics: Science, Services and Agents
on the World Wide Web, 8(4):348–354, 2010. Semantic Web
Challenge 2009 User Interaction in Semantic Web research.
[11] P. Heim, J. Ziegler, and S. Lohmann. gFacet: A browser for
the web of data. In Proceedings of the International
Workshop on Interacting with Multimedia Content in the
Social Semantic Web (IMC-SSW 2008), volume 417 of
CEUR Workshop Proceedings, pages 49–58, Aachen, 2008.
[12] A. Hogan, A. Harth, J. Umbrich, S. Kinsella, A. Polleres, and
S. Decker. Searching and browsing linked data with SWSE:
The semantic web search engine. Web Semantics: Science,
Services and Agents on the World Wide Web, 9(4):365 – 401,
2011. {JWS} special issue on Semantic Search.
[13] A. Hogan, J. Umbrich, A. Harth, R. Cyganiak, A. Polleres,
and S. Decker. An empirical survey of linked data
conformance. J. Web Sem., 14:14–44, 2012.
[14] J. M. Juran. Juran’s Quality Control Handbook.
Mcgraw-Hill (Tx), 4th edition, 1974.
[15] A. Zaveri, A. Rula, A. Maurino, R. Pietrobon, J. Lehmann,
and S. Auer. Quality assessment methodologies for linked
open data (under review). Semantic Web Journal, 2012. This
article is still under review.