=Paper=
{{Paper
|id=Vol-1704/paper7
|storemode=property
|title=VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
|pdfUrl=https://ceur-ws.org/Vol-1704/paper7.pdf
|volume=Vol-1704
|authors=Muhammad Javed,Sandy Payette,Jim Blake,Tim Worrall
|dblpUrl=https://dblp.org/rec/conf/semweb/JavedPBW16
}}
==VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation==
https://ceur-ws.org/Vol-1704/paper7.pdf
VIZ–VIVO: Towards Visualizations-driven
Linked Data Navigation
Muhammad Javed1 , Sandy Payette2 , Jim Blake3 , Tim Worrall4
Cornell University Library (CUL)
Cornell University
Ithaca, NY 14850, United States
{mj4951 , sdp62 , jeb2283 , tlw724 }@cornell.edu
Abstract. Scholars@Cornell is a new project of Cornell University Li-
brary (CUL) that provides linked data and novel visualizations of the
scholarly record. Our goal is to enable easy discovery of explicit and
latent patterns that can reveal high-impact research areas, the dynam-
ics of scholarly collaboration, and expertise of faculty and researchers.
We describe VIZ-VIVO, an extension for the VIVO framework that en-
ables end-user exploration of a scholarly knowledge-base through a con-
figurable set of data-driven visualizations. Unlike systems that provide
web pages of researcher profiles using lists and directory-style metaphors,
our work explores the power of visual metaphors for navigating a rich
semantic network of scholarly data modeled with the VIVO-ISF ontol-
ogy. We produce dynamic web pages using D3 visualizations and bridge
the user experience layer with the underlying semantic triplestore layer.
Our selection of visual metaphors enables end users to start with the big
picture of scholarship and navigate to individuals faculty and researchers
within a macro visual context. The D3-enabled interactive environment
can guide the user through a sea of scholarly data depending on the
questions the user wishes to answer. In this paper, we discuss our pro-
cess for selection, design, and development of an initial set of visualiza-
tions as well as our approach to the underlying technical architecture.
By engaging an initial set of pilot partners we are evaluating the use
of these data-driven visualizations by multiple stakeholders, including
faculty, students, librarians, administrators, and the public.
Keywords: Data Driven Documents (D3), Visualizations, Linked data,
RDF, VIVO, User Interface, Web of Data (WoD)
1 Introduction
As stewards of the scholarly record, Cornell University Library (CUL) has consis-
tently leveraged new and emerging technologies to improve access and discovery
to scholarly resources and to preserve and archive them for future generations. A
new project of CUL is Scholars@Cornell1 , a data and visualization service built
upon a semantic, linked data knowledge-base that represents the record of schol-
arship produced by Cornell faculty and researchers. One key question we set out
1
http://scholars.cornell.edu/
80
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
to answer was how can visual mediation help users navigate the rich semantic
data that represent the scholarly record? We have developed VIZ-VIVO, as an
extension to the VIVO2 framework to enable exploration of the scholarly record
through a configurable set of data-driven visualizations that leverage relation-
ships between people (e.g., faculty/researchers), their affiliations (e.g., academic
departments or colleges), and their research outputs (e.g., journal articles, books,
datasets). Our goal is to enable easy discovery of both explicit and latent pat-
terns that reveal high-impact research areas, patterns of scholarly collaboration,
and expertise of faculty and researchers.
In this paper, we discuss how we bridged the chasm between our backend of
semantically rich data in VIVO triplestore, with a fresh frontend user experience
that takes advantage of new developments in the dynamic world of web visual-
izations. We present our ongoing work on integration of D3 visualizations into
the VIVO frontend pages. The visualizations-driven approach provides efficient
overviews of the huge network of interconnected resources. D3 visualizations are
intuitive for the users to interact and offer the ability to visualize, filter and navi-
gate. The word cloud (presenting the keywords from the scholarly works) is used
not only to visualize the domain expertise of a researcher, but also to provide ac-
cess to the actual VIVO page of such scholarly work. The concept network map
and sunburst visualizations are used to present the person-to-subject-areas map
and inter-unit/cross-unit collaborations, respectively. Additionally, both tradi-
tional metrics (i.e. citation count, journal ranking.) and alternate metrics [8, 9]
are used to illustrate the impact of a particular scholarly work. By engaging
our initial of pilot partners at Cornell University, we are currently evaluating
the use of these data-driven visualizations by multiple stakeholders, including
faculty, students, librarians, administrators, and the public.
This paper is structured as follows: In Section 2 we discuss the user viewpoints
that we intended to accommodate with the VIZ-VIVO visualizations. In Section
3, we discuss how we conceived of particular visualizations that could be driven
by linked data navigation. We highlight specific examples showing sub-graphs
from the underlying VIVO data can result in particular visualizations in the
user interface layer. In Section 4, we discuss our methodology for integrating
visualizations into the VIVO framework. Our approach involved developing a
new component and API to be used with the core VIVO software and working
with users to evaluate usability and functional suitability of visualizations. A
short evaluation is given in Section 5. Related work is presented in Section 6 and
we end with some discussion.
2 VIZ–VIVO: Motivation and Viewpoints
In our work at Cornell University Library, the primary entity of interest is schol-
arship, of which people and organizations are, by definition, both the creators
and consumers. From this perspective, attention is focused on aggregate views
2
http://www.vivoweb.org
81
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
of scholarship and our visualizations become the entry points into a rich graph
of knowledge that can be explored interactively. Using VIZ-VIVO capabilities, a
user can begin with a Cornell-wide perspective of subject areas, move to views of
particular research topics, and discover faculty members in context, along with
traversing lateral pathways to collaborators across the university and beyond.
While other faculty profiling systems exist, including other VIVO-based appli-
cations and as well as services such as ResearchGate3 , these tend to use limited
list-oriented metaphors that draw the user’s attention to a primary entity –
an individual, which is presented as a profile web page. Our selection of visual
metaphors is motivated by enabling end users to start with the big picture, and
encounter individuals (e.g., faculty members, collaborators) in context, within
an interactive environment that can visually guide the user through the sea of
scholarly data depending on the questions the user wishes to answer. While users
can still view profiles or download data about individual entities, they are found
within the process of exploration when using active graph visualizations. This is
a paradigm shift in faculty profiling systems that is consistent with the mission
of the library in supporting discovery of knowledge, by starting with macro views
of scholarship upfront.
Among the user stories that inspired our Scholars@Cornell pilots, are how
the scholarly data can answer questions from different user viewpoints. For ex-
ample, how academic deans and department chairs identify where collaborations
are happening and the impact of their unit’s scholarly output? How can librar-
ians use novel views of publication patterns to inform decisions about library
holdings and preservation strategies. How can faculty and researchers have new
ways to expose their work or locate other researchers with whom they might col-
laborate. Finally, how can university executives explore macro-level overviews of
the university and its research to highlight high impact research areas and assist
them in decision making related to invest in emerging research areas. Based on
the above given viewpoints, we opt for a bottom up approach where we started
from a set of questions. Following are some specific examples of such questions
that different users would like Scholars@Cornell to answer.
– List of contributions/research output of a unit.
– List of contributions/research output of a researcher.
– List of potential collaborators (people/organizations).
– List of inter-unit/cross-unit collaborations.
– List of domain experts (for a specific subject area).
– Impact of a scholarly work.
3 Visualizations-driven Linked Data Navigation
Once user stories and the set of questions were outlined, next step was to re-
alize what data we need to answer such questions and how such data can be
presented. In regards to the presentation, we opt for Data Driven Documents
3
https://www.researchgate.net
82
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
(D3) visualization [7]. They can bridge between the best of the two worlds (i.e.
frontend and backend).
At the backend, we record data in a RDF triplestore that is modelled based
on a standard, community-driven ontology – VIVO-ISF4 . Having data in RDF
not only allowed us to create links from local entities to the external resources
but also to use the entity URIs as the local authorities, specifically for the per-
sons, publications and the organizations. Creation of such URI-driven authority
records is in alignment with the CUL’s current goals for digital preservation and
open access. Further, use of ontology-driven data modelling approach allowed us
to use inference mechanisms to realize the knowledge that is not explicitly given
in the triplestore.
A picture is worth a thousand words. At the frontend, we use D3 visual-
izations. Visualizations permit a user to navigate through the large linked data
network in an intuitive way. These are semantic-data driven visualizations that
allowed us to move from list-driven data views to visualization-driven data views
and navigation. Filters are used to narrow down the lists that consist of hun-
dreds of entries. In addition to all other benefits, visualizations also give a fresh
look and feel to the user interface.
In the following section, we discuss few of the visualizations and the questions
they address.
3.1 Person-to-Subject-Area Map (Department level )
At the department level, the concept map network (see Fig. 2) presents a map-
ping between a person and his/her research interests. Different citation indexing
applications (e.g. Web of Science, PubMed, arXiv) classify the publication venues
(e.g. journals, conferences, workshops) in different subject categories. Therefore,
all the scholarly works, published in a selected venue, receive the same category
as applied to the venue. For example, if an article is published in a journal that
is classified under the category of polymer science, the article is considered to
be related to the subject polymer science. We made use of the same transitivity
approach in order to represent the research interests of an author (see Fig. 1).
Fig. 1. Person-to-subject-area map inference
4
https://wiki.duraspace.org/display/VIVO/VIVO-ISF+Ontology
83
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
Let, journal j be the publication venue for an article t and the person p is
one of the author of the article t. If journal j is categorized under the subject
area category c, we can infer that the article t can be classified under the subject
area c and the person p has the research interest in the subject area c.
Fig. 2. Concept map network - presenting Person-to-Subject-area (P2S) map
The person-to-subject-area map is helpful for the identification of i) list of
research interests of a person and ii) list of potential collaborators. The map not
only demonstrates the overlap of research interests among different persons but
also allows a user to navigate through the underlying linked data. A profile page
link is given on every personal subject area map view. The link navigates the
user to the persons’s profile page. On the other hand, subject area views direct
a user to the subject area’s profile page, where one can learn what other entities
(e.g. book chapters, dissertations, organizations, collections) are associated to
the same subject area.
3.2 Domain Expertise of a Scholar (Person level )
The subject area terms given above in the P2S map are fairly generalized. These
terms are useful to categorize the persons (and their research publications) and
84
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
narrowing down the lists. However, in order to realize the core domain expertise
of a person, we need to take a step further.
Fig. 3. Keyword cloud - presenting domain expertise of a person
Keyword cloud (see Fig. 3) presents the domain expertise of a person. Most
of the scholarly works (such as articles, conference papers, posters, dissertations
etc.) contains a keyword section. At the person level, the cloud consists of the
keywords that are accumulated from the scholarly works being published by a
researcher. The size of each keyword (in the cloud) is directly proportional to
the number of scholarly works in which the keyword is been mentioned. Clicking
on a keyword allows a user to view the list of such scholarly works. The user can
further click on one of the work titles in order to navigate to the specific profile
page of the work and learn more details about it.
3.3 Inter-unit/Cross-unit Collaborations (College level )
The inter-unit/cross-unit collaborations are presented at the college level. At
Cornell, each college consists of a number of departments, institutes, centers
and laboratories. Though, we receive up to date persons and their positions
(affiliations) data from the human resource (HR), in order to identify the col-
laborative works, we cannot rely on this data. This is due to the following two
reasons.
– People change their affiliations with time.
– Contribution/Output of an organization do not change with joining/leaving
or change of affiliation (within organization) of a person.
For example, if a faculty member p1 (who works at an organization org1 )
co-authored an article a with a faculty member p2 (who works at organization
85
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
org2 ), article a will be considered as a collaborative work between organizations
org1 and org2 . Later, if faculty member p1 changes his/her affiliation from org1
to org3 and faculty member p2 changes his/her affiliation from org2 to org4
(where org3 and org4 can be local or external organizations), article a will still
be considered as a collaborative work between org1 and org2 .
The inter departmental and cross-unit collaborations were identified based
on the affiliations extracted from the citation data of a scholarly work. For
example, if citation data of an article describes multiple distinct affiliations for
the co-authors, then it is a collaborative work. The reconciliation process for
the affiliation data strings and the person name strings (coming from different
citation data sources) to VIVO’s person and organization URIs, is not discussed
here and is out of the scope of this paper.
Fig. 4. Sunburst visualization - presenting inter-unit/cross-unit collaborations
Sunburst visualization (see Fig. 4) is used with the tooltips in order to present
the inter-unit/cross-unit collaborations. Figure 4 presents an inter-unit/cross-
unit collaboration view for the College of Engineering (EN) 5 . The collaboration
view consists of three layers i.e., college-level layer (inner most), departmental
layer (middle) and the person-level layer (outer most). To view the person-level
layer, a user is first required to select the college of interest in the inner most
5
https://www.engineering.cornell.edu
86
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
layer. To browse the collaborations within a unit, in this example EN-to-EN, one
can select the Engineering (EN) in the college-level layer. If a user is interested
in cross-unit collaborations, s/he can select a college other than Engineering.
A user can start navigation from the college level layer and while traversing
through the department (of his/her interest), can reach to a specific person who
has co-authored some of the collaborative work. Clicking on a person’s name
section displays a tooltip that consists of the list of the (collaborative) work
titles. User can further click on one of the work titles from the list in order to
redirect to the specific profile page of such work.
4 Integration of D3 Visualizations in VIVO
Visualizations can be created in the browser using powerful JavaScript libraries
like D3 or InfoViz. Such libraries create visualizations that are animated, respon-
sive, and captivating. Server is only required to provide the data – a difficult task
in itself. We envision a range of data sets that would be required for these vi-
sualizations. This will certainly include queries against the VIVO’s triplestore
and search index. It will also include data sets that require extensive processing
and will likely be created by batch processes and cached for serving on demand.
Existing APIs in VIVO framework are not flexible enough to handle these re-
quirements. Further, the APIs do not provide the granularity of authorization
that we require. If the JavaScript on the browser can obtain data, then the same
data may be obtained by any client using HTTP, and used for any purpose.
4.1 Approach and the API Design
A data distribution API is created and is already been submitted to the VIVO
community for adoption into the core. The API permits a site administrator to
create a configuration file that enables requests for data from different sources.
The JavaScript code on the browser requests a dataset by name, supplying qual-
ifying parameters as appropriate. As each request type must be configured by
a site administrator, a user cannot make unexpected or unauthorized requests.
Requests may be restricted to logged-in users, or even to users who are logged
in with a specific authorization level. Users without sufficient authorization will
receive no data. It is important that the actual source of the data is hidden from
the JavaScript request. For example, if a request takes too long to fulfill, the site
admins can change the configuration, satisfying the request from a file of cached
data, instead of a query in real-time.
On startup, the server loads a configuration file that associates data requests
with the means of satisfying them. The configuration file includes the name of
a Java class that will supply the data, along with any parameters that the class
requires. The Java class implements a Data Distributor interface. This means
that the API controller can ask what action the class satisfies and what MIME
type should be assigned to the resulting data (see Fig. 5). The API controller
87
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
JavaScript in the browser
Event
AJAX call
HTTP request
api/dataRequest/myAction
Java on the server Data transformation
API Controller
Visualization
Configuration file
:myHandler
a DataDistributor,
MyDistributor;
:actionName "myAction".
Data Distributor
Fig. 5. The API design
asks the distributor class to produce the data and the controller routes the data
to the user.
One of the powerful concepts in this design is that a Java class can be writ-
ten as special-purpose code for a single use, or as general-purpose code which
can be configured for each use case. We expect to see more general-purpose
DataDistributor classes, such as:
– SolrDataDistributor, for queries against the search index.
– HttpRelayDataDistributor, for queries against services at other sites.
4.2 Data Transformation
We divide the workload into two tasks i.e., i) designing and implementing the
visualization and ii) providing the data. A transformation step was added to the
data flow (see Fig. 5), and a JavaScript function named loadVisualization()
was written to control the operation. The JavaScript code that creates the key-
word cloud visualization is given below as an example.
88
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
In this example, the visualization is created when the page is loaded into
the browser. The visualization loader requests the data with the action name of
person word cloud. When the data is received, the visualization loader calls the
JavaScript function transform word cloud data, that transforms a SPARQL
ResultSet to a JSON structure. Finally, the visualization loader calls the function
draw word cloud, that populates the visualization with the transformed data.
5 Evaluation
To evaluate the proposed visualizations and the linked data navigation method-
ology, journal articles, as a category of scholarly work, was selected. Citations
data for more than 11K journal articles, authored by the faculty members and
researchers of college of engineering, were uploaded in the VIVO triplestore. A
user-based evaluation approach was selected in order to empirically evaluate the
usability and the functional suitability of the integrated visualizations.
5.1 Empirical Evaluation
We involved college administrators and librarian to evaluate the visualizations.
College administrators highly ranked the zoomable collaboration view. Using
different colors for representing distinct colleges, helps a user to understand the
breakdown of collaborations between a unit and the collaborating units. The
zoom in and out functionality helps to traverse, see visualization from different
angles and to answer different questions. Such questions include “where cross-
unit collaborations are happening?”, “which departments of a college are par-
ticipating in these collaborations?”, “who are the collaborating faculty members
and how often do they collaborate?”. Further, the visualization also presents the
outcomes of these collaborations. The outcomes, such as co-authored articles,
are presented under the tooltips where a user can click on any of the publica-
tion title in order to redirect to a specific page of such scholarly output. This is
identical to the traversing of the linked data graph – from one node to the other.
Furthermore, having impact view of a scholarly work is very helpful to evaluate
the contributions and scholarly outcomes of a unit or person.
Librarians were very much attracted to the person-to-subject-area map and
the keyword cloud view. Person-to-subject-area map is helpful for understanding
the interests of a faculty member and in learning how these interests overlap
with the others. Keyword cloud view is very useful especially for those (e.g. post
graduate students) who are not only interested in learning about the research
interests of a faculty member, but also in their scholarly works. For example, if
a user has interest in the domain of chemotherapy and found this (or related)
term in the keyword cloud, by clicking on the keyword s/he can view the list of
publications, authored by the faculty member, on this topic.
89
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
6 Related Work
In last few years, a number of approaches for ontology and linked data visualiza-
tion have been emerged [2, 4–6, 10, 12, 13]. A similar work on the representation
of research expertise of a person is presented in [2]. In contrast to our method-
ology of usage of subject terms mentioned in the keyword category of an article,
author proposed a method for the keyword extraction from the text of the docu-
ment. Similar to our approach, author used word cloud view to present the expert
profiling. In [3], author presents the preliminary results on a framework for the
representation and visualization of an ontology, extracted from the text of a
document. In [4], author presents a rule-based recommendation system that was
subsequently developed to help the user choose a suitable ontology visualizer. In
[5], author presents a tool named as Visualization Playground (VISU). VISU is
a SPARQL interface that supports querying and visualizing the university data.
The goal here was to create a flexible and user friendly tool for exploring and
visualizing data from linked universities. LinkDaViz [10] is a framework that
supports the automatic binding of linked data to the visualizations. The frame-
work is based on the analysis of the input data structure and a visualization
model in order to facilitate automatic bindings. In [9], author presented a model
for assessment of research impact beyond citation analysis. Eysenbach [8] found
that the number of tweets about a research article, within the first three days of
an article’s publication, can predict which articles will be highly cited.
7 Discussion and Future Work
In the Scholars@Cornell project, we have developed VIZ-VIVO as a means of
bridging the best of the two worlds. The back-end world leverages semantic
technologies by providing an ontology-driven data model of the fundamental
entities of scholarship (people, publications, organizations). Through automated
feeds and selected curation, this data is instantiated in a VIVO triplestore to
form a graph that can be traversed, visualized, and subjected to inference. The
front-end world uses visualizations driven by sub-graphs of the entire corpus of
this scholarly data. This has enabled us to move away from user experiences
dominated by lists and directories (e.g., articles and people), towards a visual
interactive experience that leverages the power of relationships in the data.
In this paper, we presented our approach on using D3 visualizations in com-
bination with a VIVO-based semantic knowledge base of the scholarly records of
Cornell University. Our work enables easy discovery of subject area and domain
expertise of faculty members and researchers, as well as the discovery of patterns
in the scholarly records. Our approach offers a mean of presenting linked data
through dynamic web pages with visualizations that permit a user to navigate
a large corpus of semantic data and visually explore interdisciplinary collabora-
tions and emerging trends in scientific publications.
In the future, we will be working with users to explore and evaluate other
types of visualizations based on the more complex questions they seek to answer
90
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
in VIVO. We will also evolve our data to express common categorization schemes
for different disciplines to better support visualizations of research areas. Our
user interface will evolve to make all data that underlies each visualization down-
loadable in common formats such as csv, json, and xml. Our continuing work
will focus on making linked data easily discoverable, consumable, visualizable,
and downloadable to bring the power of semantic data to users of VIVO and
other linked data applications.
Acknowledgements. Other members of the Scholars@Cornell team contributed
to backend systems development, data modeling, data feeds, and data curation.
We especially acknowledge the efforts of Joe McEnerney, Jill Wilson, Jason Ko-
vari, and George Kozak.
References
1. Krafft, D.B., Cappadona, N.A., Caruso, Corson-Rikert, J., Devare, M., Lowe, B.J.,
& VIVO Collaboration. VIVO: Enabling national networking of scientists. Paper
presented at the Web of Science Conference, raleigh, NC, 2010.
2. Koperwas, J., Skonieczny, L., Kozlowski, M., Andruszkiewicz, P., Rybiński, H.,
Struk, W.: Intelligent information processing for building university knowledge base.
Journal of Intelligent Information Systems, Springer, pages 1–23, 2016.
3. Dasiopoulou, S., Lohmann, S., Codina, J., Wanner, L.: Representing and visualizing
text as ontologies: A case from the patent domain.In: Proceedings of the Interna-
tional Workshop on Visualizations and User Interfaces for Ontologies and Linked
Data (VOILA 2015). CEUR-WS, vol. 1456, pp. 83–90 (2015)
4. Dudáš, M., Zamazal, O., Svátek, V.: Roadmapping and navigating in the ontology
visualization landscape. In 19th International Conference on Knowledge Engineering
and Knowledge Management, pages 137–152. Springer, 2014.
5. Alonen, M., Kauppinen, T., Suominen, O., Hyvonen, e.: Exploring the Linked Uni-
versity Data with Visualization Tools. In proceedings of the European Semantic
Web Conference (ESWC), 2013.
6. Atemezing G.A., Troncy, R.: Towards a linked-data based visualization wizard. In
Proceedings of the 5th International Workshop on Consuming Linked Data (COLD
2014) co-located with the ISWC, 2014.
7. Bostock, M., Ogievetsky, V., Heer, J.: D3 data-driven documents. IEEE Transac-
tions on Visualization and Computer Graphics (TVCG). volume 17, number 12,
pages 2301 – 2309, 2011.
8. Eysenbach, G.: Can tweets predict citations? Metrics of social impact based on
Twitter and correlation with traditional metrics of scientific impact. Journal of
Medical Internet Research (JMIR), volume 13, issue number 4, 2011.
9. Sarli, C.C., Dubinsky, E.K., Holmes, K.L.: Beyond citation analysis: a model for
assessment of research impact. Journal of the Medical Library Association (JMLA),
volume 98, issue number 1, pages 17–23, 2010.
10. Thellmann, K., Galkin, M., Orlandi, F., Auer, S.: LinkDaViz – Automatic Binding
of Linked Data to Visualizations In proceedings of International Semantic Web
Conference (ISWC), 2015.
91
VIZ-VIVO: Towards Visualizations-driven Linked Data Navigation
11. Abello, J., van Ham, F., Krishnan, N.:. ASK-GraphView: A Large Scale Graph
Visualization System. IEEE Transactions on Visualization and Computer Graphics
(TVCG), volume 12, number 5, 2006.
12. Bikakis, N., Liagouris, J., Krommyda, M., Papastefanatos, G., Sellis, T.: Towards
Scalable Visual Exploration of Very Large RDF Graphs. In proceedings of the Eu-
ropean Semantic Web Conference (ESWC), 2015.
13. Frischmuth, P., Martin, M., Tramp, S., Riechert, T., Auer, S.: Ontowiki–an author-
ing, publication and visualization interface for the data web. Semantic Web Journal,
volume 6, number 3, pages 215–240, 2015.
14. Tang, J., Zhang, J., Yao, L., Li, J., Zhang, L. and Su, Z.: Arnetminer: extraction
and mining of academic social networks. In Proceedings of the 14th ACM SIGKDD
International Conference on Knowledge Discovery and Data mining. pages 990–998,
2008.
15. Osborne, F., Motta, E. and Mulholland, P.: Exploring scholarly data with Rexplore.
In International Semantic Web Conference (ISWC), pages 460–477, Springer Berlin
Heidelberg, 2013.
16. Monaghan, F., Bordea, G., Samp, K. and Buitelaar, P.: Exploring your research:
Sprinkling some saffron on semantic web dog food. In Semantic Web Challenge at
the International Semantic Web Conference (ISWC), vol. 117, pages 420–435, 2010.
92