=Paper=
{{Paper
|id=Vol-2778/paper1
|storemode=property
|title=A Comparative Study of State-of-The-Art Linked Data Visualization Tools
|pdfUrl=https://ceur-ws.org/Vol-2778/paper1.pdf
|volume=Vol-2778
|authors=Federico Desimoni,Nikos Bikakis,Laura Po,George Papastefanatos
|dblpUrl=https://dblp.org/rec/conf/semweb/DesimoniBPP20
}}
==A Comparative Study of State-of-The-Art Linked Data Visualization Tools==
https://ceur-ws.org/Vol-2778/paper1.pdf
A Comparative Study of State-of-The-Art
Linked Data Visualization Tools
Federico Desimoni1[0000−0002−9051−2741] , Nikos Bikakis2[0000−0001−6859−1941] ,
Laura Po1[0000−0002−3345−176X] , and George
Papastefanatos2[0000−0002−9273−9843]
1
University of Modena and Reggio Emilia, Italy
{fdesimoni,laurapo}@unimore.it
2
ATHENA Research Center, Greece
{bikakis,gpapas}@athenarc.gr
Abstract. Data visualization tools are of great importance for the ex-
ploration and the analysis of Linked Data (LD) datasets. Such tools
allow users to get an overview, understand content, and discover inter-
esting insights of a dataset. Visualization approaches vary according to
the domain, the type of data, the task that the user is trying to perform,
as well as the skills of the user. Thus, the study of the capabilities that
each approach offers is crucial in supporting users to select the proper
tool/technique based on their need.
In this paper we present a comparative study of the state-of-the-art LD
visualization tools over a list of fundamental use cases. First, we define 16
use cases that are representative in the setting of LD visual exploration,
examining several tool’s aspects; e.g., functionality capabilities, feature
richness. Then, we evaluate these use cases over 10 LD visualization
tools, examining: (1) if the tools have the required functionality for the
tasks; and (2) if they allow the successful completion of the tasks over
the DBpedia dataset. Finally, we discuss the insights derived from the
evaluation, and we point out possible future directions.
Keywords: Visualization Use Cases · Usability · Tools Benchmark · Ontology
Visualization · RDF Graph Data · OWL · Linked Data · Semantic Web
1 Introduction
The wide adoption and availability of a very large number of Linked Data (LD)
sources makes data visualization and exploration a crucial task for most LD
consumers. Data scientists, domain experts, business users and citizens, wish to
employ intuitive and visual, rather than programmatic, ways to interact with
the data.
Copyright 2020 for this paper by its authors. Use permitted under Creative Commons License
Attribution 4.0 International (CC BY 4.0).
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
The exploration of LD is a particular task that differs from the classical data
visualization, mainly due to the LD characteristics. The use of common (usually
cross-domain) vocabularies for the description of the data (e.g., a resource is an
instance of a class) or the use of typed properties for capturing relationships
between resources within a dataset or across datasets, makes traditional data
visualization ways, such as bar charts, incapable of capturing the complex inter-
relationships within a LD dataset. Graph-based or similar approaches are usually
employed. In addition, users, usually accessing a remote SPARQL endpoint, do
not have an a-priori knowledge of the dataset, do not know if the dataset might
be relevant for their goals and follow an exploratory way to visualize informa-
tion. Finally, visualizing LD means to handle several issues: the large size, and
the dynamic nature of data, the requests for exploratory searches, the variety of
tasks, and the different types and needs of users.
A large number of LD visualization tools have been introduced the recent
years, most of them originating from the academic sector. LD visualization tools
provide graphical representations of a dataset or parts of it, with the aim of
facilitating their analysis and generating insights into complex interconnected
information. Visualization techniques can vary according to the domain, the
type of data, the task that the user is trying to perform, as well as the skills
of the user. To this end, the evaluation of the functionality that each technique
offers, is necessary for users to better fit a tool/technique to their need.
Visual methods for query formulation undertake the challenge of making
querying independent of users’ technical skills [29]. There is a plethora of works
that attempt to provide systematic reviews and surveys for several aspects of
LD exploration and visualization. The recently published LD visualization tools
book [24] includes an extensive review of such tools. Also, there are other surveys
[10,7,17,21,1,5], which also address different aspects of visualization methods
and tools. For example, [17,3] studies issues related to the entire process of LD
consumption and exploration.
In contrast, in this study, we evaluate the functionality capabilities and fea-
tures richness of several tools over a list of use cases, which are fundamental
in the context of LD visual exploration. We define the use cases and evaluate
whether each tool enables users to perform each use case over the DBpedia
dataset. Our evaluation studies the following two aspects, if the tools: (1) pro-
vide the functionality/feature that is required to perform the tasks; and (2) allow
the successful completion of the tasks over a dataset.
Contribution. (1) We define 16 use cases related to LD visualization, which
allow us to evaluate several functionality aspects of LD visualizations tools. The
use cases cover a plethora of tasks, focusing on different visualization aspect. For
example, use cases related to: core LD concepts (e.g., schema, classes, instances,
properties), concepts relationships (e.g., paths, common properties), navigation,
visualization recommendation, filtering, statistics, etc. (2) We examine a long
list of tools, and we select 10 tools which are publicly accessible and available
for use. (3) We evaluate the tools by examining the defined use cases over the
2
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
DBpedia dataset. The use of DBpedia allows to study the tools over real-world
tasks and scenarios.
2 An Overview of the Evaluated Tools
This section provides a brief description of the tools used in our study. Regarding
the tools’ selection process, we initially consider the list of the 29 tools presented
in [11,24], then, we selected the tools that are publicly accessible, available for
use, and support SPARQL endpoint access. As a result, 10 tools are selected, 9
out of 10 are available online as web services, while Tarsier is an open source
tool, which its code is available on GitHub. More details about the tools could
be found at [11,24].
RelFinder 3 [18] is a Web-based tool that offers interactive discovery and
visualization of associations (i.e., relationships) between selected LD instances.
In this context, as associations are considered the paths (i.e., set of properties)
that connect the instances in the LD graph.
LinkedOpenGraph 4 (LOG) [4] visualizes LD following a graph node-link lay-
out. LOG allows the users to explore the visualized LD graph, by offering several
interactive functionalities, such as, zoom, pan, filter, keyword search, and edit
over nodes and edges (e.g., delete, rename, change color and shape, rearrange
layout). Similarly, Tarsier 5 [31] tool visualizes LD using 3D node-link graph
representation.
LodLive 6 [8] and LodView 7 are graph-based tools that adopt the incremental
visualization paradigm, in which, instead of the whole graph, a starting point
(e.g., node) is visualized, then, based on user interaction, more part of the graph
are presented. In this context, the user starts her exploration from a given URI
or a SPARQL endpoint.
LD-VOWL8 [32] uses SPARQL queries to process RDF triples in order to
infer schema information. The tool first identifies and presents the most repre-
sentative concepts, using several methods and assumptions, (e.g., consider the
classes with larger number of instances as representative). Then, the ontology
schema is (progressively) visualized as graph, offering several interactive oper-
ations. The schema/ontology visualizations generated by the LD-VOWL follow
the VOWL [20,19] graphical representation.
H-BOLD 9 (High-level visualization over Big Open Linked Data) [23,25] gen-
erates a representative summary of a LD source. BOLD takes as input a SPARQL
endpoint and generates a visual (graph-based) summary of the LD source, ac-
companied by statistical and structural information of the source. Further, BOLD
3
http://www.visualdataweb.org/relfinder/relfinder.php
4
https://log.disit.org/service
5
https://github.com/desmovalvo/tarsier
6
http://lodlive.it
7
https://lodview.it
8
http://vowl.visualdataweb.org/ldvowl
9
https://dbgroup.ing.unimo.it/hbold
3
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
offers incremental multilevel exploration, where a community detection algo-
rithm is used to effectively construct the abstract levels.
In order to assist and guide the users in visual exploration scenarios, LD-
VizWiz 10 (Linked Data Visualization Wizard) [2] exploits data semantics to
simplify the process of setting up visualizations, providing a semi-automatic
way for the production of possible visualization for LD datasets. LDVizWiz of-
fers several visualizations types, i.e., Map, Tree/Hierarchy and Pie.
RDFSurveyor 11 [30] is a LD browser that provides class-based navigation
over data retrieved from SPARQL endpoints. This tool is similar to a traditional
Web browser, allowing the user to navigate over the dataset throw links that rep-
resent LD properties. Similarly, SPARKLIS 12 [12] is a LD browser that combines
the faceted exploration paradigm (a.k.a. faceted search) with query building.
Based on user’s selections, the tool incrementally build structure queries, based
on which the explored data is selected.
Note that, there are several tools that are available for use and cover nu-
merous of the use cases evaluated in this paper; e.g., Balloon Synopsis [28],
graphVizdb [6], LODmilla [22], Phuzzy.link [26], RDF4U [9], Rhizomer [13],
QueryVOWL [15]. However, these tools are not considered in our evaluation
since do not support SPARQL endpoint access.
3 Linked Data Visualization Use Cases
Users perform specific tasks while visually exploring LD and these tasks can be
rendered into features implemented by visualization tools. The work in [24] has
identified and categorized the possible interactions between a tool and a user
into a number of use cases (UC). A use case describes how a user makes use of
a system to accomplish a particular goal [16]. Use cases capture common func-
tionalities that LD visualization tools should support, for users to gain clear and
convincing visualization of the information (i.e., schema, resources and statis-
tics), contained in a LD source. Detailed examples about the use cases can be
found in [24].
Based on this list, we employ for our evaluation purposes 16 use cases, con-
sidering also generalization and special cases of some of the use cases included
in [24]. The employed use cases, cover a plethora of tasks, which allow us to
examine several tasks related to schema-level (e.g., classes, properties) and data-
level (e.g., class’ instances, values of properties) visualization, and the relation-
ships between the visualized concepts; e.g., paths, common properties values,
etc. Moreover, tasks related to exploration capabilities for navigation, visualiza-
tion recommendation, filtering, and statistics presentation. The examined use
cases, with more details on their intended functionalities are:
10
http://semantics.eurecom.fr/datalift/rdfViz/apps
11
http://tools.sirius-labs.no/rdfsurveyor
12
http://www.irisa.fr/LIS/ferre/sparklis
4
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
UC1 Visualize OWL/RDF Schema. User selects a dataset to visualize
schema information; e.g., classes, properties, classes/properties rela-
tions (e.g., super, equivalent, disjoint), semantics (e.g., inverse prop-
erty, carnality), etc.
This information useful for obtaining a global view of the dataset’s
schema. User can understand with a glimpse which classes the dataset
is composed by, which properties connect each of them, and several
semantics.
– UC1-C Visualize OWL/RDF Schema classes. User selects a dataset
to visualize classes’ information. This is a special case of UC1, in which
the information refers only to classes.
– UC1-P Visualize OWL/RDF Schema properties. User selects a
dataset to visualize (object, datatype) properties’ information. This use
case can be considered as a special case of UC1, in which the visualized
information refers only to properties.
UC2 Visualize the information related to a class. User selects a specific
class to visualize class’s information This use case can be considered
as a special of UC1-C, in which the visualized information refers to a
specific class.
For large T-Boxes it is useful to analyze single elements instead of the
complete schema. For example, a users may be interested in discovering
the different kind of relations connecting the selected class to other
classes (e.g., sub-class, disjoint), etc.
UC3 Visualize the information related to a property. User selects a
property to visualize property’s information. This use case can be con-
sidered as a special of UC1-P, in which the visualized information refers
to a specific property.
For example, a user may be interested in examining the characteris-
tics of a property, such as: domain, range, carnality, relations to other
properties (e.g., inverse, equivalent), etc.
UC4 Visualize instances. User selects a dataset to visualize its instances.
This use case is applicable for small LD datasets, allowing the user to
examine data-level information; e.g., relations between the instances.
UC5 Visualize the information related to a specific instance. User se-
lects an instance to visualize its object and datatype properties’ values.
Following UC4, a user may wish to focus her attention on specific
instances analyzing their properties and their relations.
UC6 Visualize the instances belonging to a class. User selects a class
to visualize its instances.
UC7 Visualize the paths that connect different instances. User se-
lects a set of instances to visualize the paths (i.e., graph paths -set of
properties-) among them.
5
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
This provides useful information regarding the direct or indirect ways
that different instances are related.
UC8 Navigate by traversing dataset’s paths. User selects a class, or an
instance to explore its properties (i.e., paths).
For example, in a node-link graph layout, the presented part of the
graph could be extended by clicking on an instance/class property (i.e.,
progressive visualization).
UC9 Provide information about the vocabularies that are used in
the dataset. User selects a dataset to examine the vocabularies used.
Vocabularies could provide useful information related to the prove-
nance of the terms used in the dataset.
UC10 Provide statistics about the dataset. User selects a dataset to an-
alyze various statistics; e.g., instances per class, frequencies.
UC11 Visualize the data by selecting visualization types (e.g., map,
plot) that are determined by the data contents. User selects a
dataset, generate or recommend different visualization types based on
the types and semantics of data; e.g., instance type and properties,
properties’ values, class hierarchies.
For example, a tool generates a map visualization, for the instances
having the geo:lat and geo:long attributes, or a treemap for class hier-
archies.
UC12 Visualize the instances that have specific properties. User se-
lects a set of properties to visualize the instances that have properties
contained in the given properties’ set.
UC13 Visualize type-specific properties of an instance. User selects an
instance and a set of properties to visualize information regarding the
given properties.
UC14 Visualize range-specific datatype properties of an instance. User
selects an instance and a a datatype (e.g., integer) to visualize infor-
mation of properties for which their range is of the specific datatype.
It is a common case, where instances have a long list of datatype prop-
erties. In such cases, filtering datatype properties based on its range
datatype could help users to identify the information they are looking
for.
UC15 Visualize range-specific object properties of an instance. User
selects an instance and a class, to visualize information regarding the
object properties which their range contained in the given class.
Similar to the UC14, a user could be interested in visualizing properties
with a specific range object.
UC16 Visualize more than one instances in parallel. User selects multi-
ple instances to visually compare information (e.g., properties values)
6
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
Use Case Selected Resources �
UC2 Book
UC3 series
UC5 The Lord of the Rings
UC6 Book
UC7 The Lord of the Rings & Manchester
UC8 The Lord of the Rings & Manchester
UC9 The Lord of the Rings & Manchester
UC12 series
UC13 The Lord of the Rings & author
UC14 The Lord of the Rings & xsd:dateTime
UC15 The Lord of the Rings & The Hobbit
UC16 The Lord of the Rings & Manchester
Table 1: DBpedia Resources Used in each Use Case
�
Book: class series, author: object properties
The Lord of the Rings, Manchester, The Hobbit: instances
related to these instances. Particularly, this use case can be considered
as a generalization of the UC5, where more than two instances are
visualized.
In this use case, a user is able to compare and identify possible com-
monalities (e.g., same property values), relations (e.g., have a path
connection), or dissimilarities between the selected instances.
4 Evaluation
In this Section, we present the evaluation of the 10 tools presented in Section
2 w.r.t. the 16 use cases described in Section 3. For each tool and use case,
we examine whether the tool gives the ability to the user to accomplish the
respective task over the DBpedia dataset.
4.1 Setup
Dataset. In our evaluation, we use the well-known, generic DBpedia dataset,
through its SPARQL endpoint13 . At the time of our evaluation (August 2020),
DBpedia endpoint accesses about 1.6B triples.
13
http://dbpedia.org/sparql
7
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
Use Cases over the DBpedia Dataset. In order to define the proposed use
cases as specific tasks over the DBpedia, we use the following resources (i.e.,
classes, properties, instances) as described in the DBpedia dataset’s description
page14 .
For class-related UCs, we consider the class Book, which has 41 outcoming
and 5 incoming properties, and about 64K instances. We also consider the object
properties series and author. The series property has the owl:Thing as domain
and range, and connects resources to the series they belong to (e.g., TV se-
ries, Book series). For example, the resource Friends is the series of the episode
The One with the Proposal. The author property has as domain the class Work
and range the class Person.
Finally, we consider the following three instances: (1) The Lord of the Rings
which is an instance of the class Book, having 480 outcoming and 1.4K incoming
properties; (2) Manchester which is an instance of the class City, having about
1K outcoming and 19K incoming properties; and (3) The Hobbit which is an
instance of the class Work.
The use of the resources per use case is presented in Table 1. For example,
the UC6 is expressed as “Visualize all the instances of the class Book”, whereas
the UC7 is expressed as “Visualize the paths (i.e., set of properties) between the
The Lord of the Rings and the Manchester”.
Evaluation Setting. The evaluation has been conducted on a laptop with i7
CPU at 2.0GHz, 16G RAM and 15.6” (1366×768) screen, running Windows 10
and Google Chrome 84.0.4147. In the evaluation, the online versions of the 10
tools are used (URL are presented in Section 2). Tarsier is installed and evaluated
locally on the laptop.
4.2 Results
Table 2 summarizes the results of our evaluation. The table indicates if a tool
has the functionality to perform a use case, as well as if the tool was able to
successfully complete a use case over the DBpedia.
UC1- C & UC1-P. Regarding the visualization of OWL/RDF schema infor-
mation, H-BOLD and LD-VOWL provide the required functionality. H-BOLD
allows users to understand the schema in a stepwise approach by showing a
summarized version of it and then presents more elements incrementally. LD-
VOWL, initially presents the complete schema and provides functionality that
allows users to hide undesired elements and focus on the interesting ones. How-
ever, in our evaluation, in both tools, we are not able to perform schema-related
tasks. Particularly, in the UC1-C and UC1-P use cases, after a long time, both
tools are not responding.
Regarding the tools that follow a tabular-like data representation, Sparklis is
able to successfully complete both use cases. Particularly, it presents classes and
properties in two different list, where the user can select the resource of interest
14
https://wiki.dbpedia.org/data-set-38
8
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
Use Case
Tool 1C 1P 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
H-BOLD � � � � � �
LD-VizWiz ○ �
LD-VOWL � � �
LodLive ○ ○ ○ ○ ○ ○
LodView ○ ○ ○ ○ ○ ○
LOG ○ ○ ○ ○ ○ ○ ○
RDFSurveyor ○ ○ ○ ○ ○ ○
RelFinder ○
Sparklis ○ ○ ○ ○ ○
Tarsier ○ ○ ○ ○ ○ ○
Table 2: Evaluation Results
.
� use case supported ○ use case accomplished
in order to further explore it. Additionally, RDFSurveyor accomplishes only the
UC1-C, since it is able to display only a list of classes, without presenting any
information about the properties. Finally, the rest of the tools do not provide
the functionality for UC1-C and UC1-P use cases.
UC2, UC3 & UC5. Regarding the use cases that request information about
a class (UC2), a property (UC3), and an instance (UC5), we report on the
following.
LodLive, LodView, LOG and Tarsier are able to successfully complete all
of the use cases. Particularly, in LodLive, LodView and LOG the user is able
to retrieve information about a resource by giving its URI. In Tarsier, this is
accomplished through SPARQL queries that are expressed by the user and eval-
uated over an endpoint. RDFSurveyor is able to perform UC5, in which the user
is able to navigate over the hierarchy of classes in order to indicate the required
instance. Note that, H-BOLD has this functionality, however, in our evaluation,
we are not able to use it (see UC-1C & UC1-P). The rest of the tools do not
provide the functionality required to perform any of these use cases.
9
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
UC4. Sparklis is the only tool that is able to accomplish UC4 by presenting
a list of instances included in a dataset. Initially, a list of about 200 instances
are presented; then, the user is able to use the search functionality in order to
find the resource of interest. The rest of the tools do not provide the required
functionality for this use case.
UC6. LOG, LodView, RDFSurveyor, and Sparklis are able to visualize the
instances of a selected DBpedia class. Particularly, LOG follows a node-link
layout, in which a class’ instances can be visualized using the incoming rdf:type
class’ properties. LodView, RDFSurveyor, and Sparklis present data in a tabular
form, in which the (web browser’s) URI links of the incoming rdf:type properties
can be used to explore instances. Note that, H-BOLD has this functionality,
however, in our evaluation, we are not able to use it (see UC-1C & UC1-P).
UC7. RelFinder offers discovery and visualization of relations between instances.
Hence, we can claim that RelFinder is mainly designed to support UC7. In our
evaluation, RelFinder is the only tool that supports UC7.
UC8. In this use case we examine which tools are able to navigate over a
dataset through properties, starting from a class or an instance. In this use case,
beyond RelFinder, four tools are able to navigate through paths; i.e., LOG,
LodLive, LodView, and RDFSurveyor. Particularly, LOG and LodLive visualize
the dataset adopting a node-link layout. They follow a progressive visualization
approach, in which, new parts of the graphs are presented, each time the user
click on a node. This way, the user can navigate following the nodes’ paths.
LodView and RDFSurveyor present the information in a tabular form. In this
case, the user are able to navigate through URI web browser’s links.
UC9. Regarding providing vocabularies information, RDFSurveyor presents the
vocabularies used in the dataset. H-BOLD and LD-VOWL have similar features,
however, in our evaluation, we were not able to use these tools over the DBpedia
data (see UC-1C & UC1-P).
UC10. This use case examines the capability of the tools to provide statistical
information about the dataset. H-BOLD provides basic statistics regarding the
datasets; e.g., the number of classes, instances, triples, properties, etc. Addition-
ally, LD-VizWiz reports information regarding the types of the data contained
in the dataset, as well as extracts some of the most common elements.
UC11. Regarding the capability of the tools to generate different visualization
types based on the types of data, LDVizWiz exploits data semantics and gen-
erate/recommend different visualization types for the data. LodLive, LodView,
and RDFSurveyor are able to identify the properties geo:lat and geo:long and
present a map with instances’ locations.
UC12. Sparklis is the only tool able to filter the instances based on a specific
property.
UC13, UC14 & UC15. As described in UC 3, LodView, LodLive, LOG and
Tarsier can present information for a particular property. In UC13 we want the
10
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
tools to present a set of specific instance’s properties. LodLive, LOG and Tarsier
are able to perform such tasks, using a filtering functionality which allows to
filter out properties. There are no tool to support UC14, in which datatype
properties are filtered based on its range datatype. Regarding the UC15, using
the filtering functionality of Tarsier we are able to indirectly accomplish UC15.
UC16. This use case can be accomplished only by the LOG and the Tarsier
tools. LOG allows the user to select and visualize multiple instances from an
endpoint, by giving its URI. In Tarsier, this is accomplished through SPARQL
queries that are expressed by the user and evaluated over an endpoint.
5 Discussion
The results presented in the previous section (Table 2) reveal the heterogeneity
of the tools in accomplishing different user needs. Tools handle tasks by follow-
ing different methods and/or workloads, or by adopting different visualization
techniques. Generally, the adopted approaches vary according to the domain, the
type of data, the task that the user is trying to perform, as well as the skills of the
user. The evaluation’s findings allow users to select the proper tool/technique
based on their need. For example, we can claim the following.
If the users goal is the visualization of the structure/schema of the dataset,
then, H-BOLD and LD-VOWL can be used. H-BOLD shows a summarized ver-
sion of the data, and then present more elements incrementally, while LD-VOWL
presents the complete schema. Such tasks can be also accomplished by the RDF-
Surveyor and Sparklis, which provide a tabular representation of the data.
If the user wishes to obtain a statistical overview of a dataset, H-BOLD or
LDVizWiz can be an option. H-BOLD provides basic statistics regarding the
datasets (e.g., the number of classes, instances); and LDVizWiz can extract the
most common elements, as well as other statistic-related information.
LodLive and LOG may be the most suitable tools for visualizing instances
and their relations. They offer specific features related to owl: sameAs relations
and inverse properties. In addition, LOG allows to insert more elements on the
screen and is able to handle and compare many datasets simultaneously.
Finally, If the user wishes to explore the relations between two or more in-
stances, the RelFinder can effectively support such tasks.
6 Conclusion and Future Work
In this paper, we evaluated and compared the functionality capabilities and
features richness of several Linked Data Visualization tools over a list of use
cases. The evaluation has been conducted over the DBpedia dataset.
We are aware that this is only a first step of a deeper analysis of these tools.
In the near future, we intend to include a user evaluation about effectiveness and
efficiency of each tool in performing the defined use cases, following the metrics
presented in [27]. Alternatively, if involving users is not feasible, we will explore
11
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
the use of usability evaluation frameworks that do not require users involvement
but provide similar metrics, like [14].
Acknowledgment. This work is partially funded by the project VisualFacts
(#1614 - 1st Call of the Hellenic Foundation for Research and Innovation Re-
search Projects for the support of post-doctoral researchers); and by the “Enzo
Ferrari” Engineering Department of the University of Modena and Reggio Emilia
(within “Networking on Linked Data” project) and by the Connection Europe
Facility of the European Union (within “TRAFAIR Understanding traffic flows
to improve air quality” project [AGREEMENT No INEA/CEF/ICT/A2017/
1566782]).
References
1. Alahmari, F., Thom, J.A., Magee, L., Wong, W.: Evaluating Semantic Browsers
for Consuming Linked Data. In: Australasian Database Conference, ADC (2012)
2. Atemezing, G.A., Troncy, R.: Towards a Linked-Data Based Visualization Wizard.
In: International Workshop on Consuming Linked Data (COLD) (2014)
3. Bauer, F., Kaltenbapck, M.: Linked Open Data: The Essentials: A Quick Start
Guide for Decision Makers. Edition mono/monochrom (2012)
4. Bellini, P., Nesi, P., Venturi, A.: Linked open graph: Browsing multiple SPARQL
entry points to build your own LOD views. J. Vis. Lang. Comput. 25(6) (2014)
5. Bikakis, N.: Big Data Visualization Tools. In: Encyclopedia of Big Data Technolo-
gies, 2nd Edition. Springer (2021)
6. Bikakis, N., Liagouris, J., Krommyda, M., Papastefanatos, G., Sellis, T.K.:
GraphVizdb: A Scalable Platform for Interactive Large Graph Visualization. In:
ICDE. pp. 1342–1345 (2016)
7. Bikakis, N., Sellis, T.K.: Exploration and Visualization in the Web of Big Linked
Data: A Survey of the State of the Art. In: EDBT (2016)
8. Camarda, D.V., Mazzini, S., Antonuccio, A.: LodLive, exploring the web of data.
In: I-SEMANTICS (2012)
9. Chawuthai, R., Takeda, H.: RDF Graph Visualization by Interpreting Linked Data
As Knowledge. In: JIST (2015)
10. Dadzie, A., Pietriga, E.: Visualisation of Linked Data - Reprise. Semantic Web
Journal 8(1) (2017)
11. Desimoni, F., Po, L.: Empirical evaluation of Linked Data visualization tools. Fu-
ture Generation Computer Systems 112 (2020)
12. Ferré, S.: SPARKLIS: a SPARQL Endpoint Explorer for Expressive Question An-
swering. In: International Semantic Web Conference (Posters & Demos) (2014)
13. Garcı́a, R., Gil, R.: Building a Semantic Intraweb with Rhizomer and a Wiki. In:
IntraWebs Workshop, World Wide Web Conference (2006)
14. Gonzlez, R.G., Iranzo, R.M.G., Bakke, E., Karger, D.R.: A benchmark for end-
user structured data exploration and search user interfaces. Journal of Web
Semantics 65(8), 100610 (2020). https://doi.org/10.1016/j.websem.2020.100610,
https://doi.org/10.1016/j.websem.2020.100610
15. Haag, F., Lohmann, S., Siek, S., Ertl, T.: QueryVOWL: Visual Composition of
SPARQL Queries. In: ESWC. vol. 9341, pp. 62–66 (2015)
16. Kim Hamilton, R.M.: Learning UML 2.0. O’Reilly (2006)
12
A Comparative Study of State-of-The-Art Linked Data Visualization Tools
17. Klı́mek, J., Skoda, P., Necaský, M.: Survey of Tools for Linked Data Consumption.
Semantic Web Journal (2018)
18. Lohmann, S., Heim, P., Stegemann, T., Ziegler, J.: The RelFinder user interface:
interactive exploration of relationships between objects of interest. In: IUI (2010)
19. Lohmann, S., Negru, S., Haag, F., Ertl, T.: VOWL 2: User-Oriented Visualization
of Ontologies. In: EKAW (2014)
20. Lohmann, S., Negru, S., Haag, F., Ertl, T.: Visualizing Ontologies with VOWL.
Semantic Web Journal 7(4) (2016)
21. Marie, N., Gandon, F.L.: Survey of Linked Data Based Exploration Systems. In:
International Workshop on Intelligent Exploration of Semantic Data (IESD) (2014)
22. Micsik, A., Turbucz, S., Györök, A.: LODmilla: A Linked Data Browser for All.
In: SEMANTiCS. pp. 31–34 (2014)
23. Po, L.: High-level Visualization Over Big Linked Data. In: International Semantic
Web Conference (P&D/Industry/BlueSky) (2018)
24. Po, L., Bikakis, N., Desimoni, F., Papastefanatos, G.: Linked Data Visualization:
Techniques, Tools, and Big Data. Synthesis Lectures on the Data, Semantics, and
Knowledge, Morgan and Claypool (2020)
25. Po, L., Malvezzi, D.: Community detection applied on big linked data. J.
UCS 24(11), 1627–1650 (2018), http://www.jucs.org/jucs_24_11/community_
detection_applied_on
26. Regalia, B., Janowicz, K., Mai, G.: Phuzzy.link: A SPARQL-powered Client-Sided
Extensible Semantic Web Browser. In: VOILA. pp. 34–44 (2017)
27. Sánchez, J.L.G., Garcı́a, R., Brunetti, J.M., Gil, R., Gimeno, J.M.: Using SWET-
QUM to compare the quality in use of semantic web exploration tools. J. Univers.
Comput. Sci. 19(8), 1025–1045 (2013). https://doi.org/10.3217/jucs-019-08-1025,
https://doi.org/10.3217/jucs-019-08-1025
28. Schlegel, K., Weißgerber, T., Stegmaier, F., Seifert, C., Granitzer, M., Kosch, H.:
Balloon Synopsis: A Modern Node-centric Rdf Viewer and Browser for the Web.
In: ESWC (2014)
29. Soylu, A., Kharlamov, E., Zheleznyakov, D., Jiménez-Ruiz, E., Giese, M., Skjæve-
land, M.G., Hovland, D., Schlatte, R., Brandt, S., Lie, H., Horrocks, I.: Op-
tiquevqs: A visual query system over ontologies for industry. Semantic Web 9(5),
627–660 (2018). https://doi.org/10.3233/SW-180293, https://doi.org/10.3233/
SW-180293
30. Vega-Gorgojo, G., Slaughter, L., von Zernichow, B.M., Nikolov, N., Roman, D.:
Linked Data Exploration With RDF Surveyor. IEEE Access 7 (2019)
31. Viola, F., Roffia, L., Antoniazzi, F., D’Elia, A., Aguzzi, C., Cinotti, T.S.: Interactive
3D Exploration of RDF Graphs through Semantic Planes. Future Internet 10(8)
(2018)
32. Weise, M., Lohmann, S., Haag, F.: LD-VOWL: Extracting and Visualizing Schema
Information for Linked Data Endpoints. In: VOILA@ISWC (2016)
13