=Paper=
{{Paper
|id=Vol-1264/AtemezingT
|storemode=property
|title=Towards a Linked-Data based Visualization
Wizard
|pdfUrl=https://ceur-ws.org/Vol-1264/cold2014_AtemezingT.pdf
|volume=Vol-1264
|dblpUrl=https://dblp.org/rec/conf/semweb/AtemezingT14
}}
==Towards a Linked-Data based Visualization
Wizard==
https://ceur-ws.org/Vol-1264/cold2014_AtemezingT.pdf
Towards a Linked-Data based Visualization
Wizard
Ghislain Auguste Atemezing1 , Raphaël Troncy1
EURECOM, Campus SophiaTech, France
Abstract. Datasets published in the LOD cloud are recommended to
follow some best practice in order to be 4-5 stars Linked Data compli-
ant. They can often be consumed and accessed by different means such
as API access, bulk download or as linked data fragments, but most of
the time, a SPARQL endpoint is also provided. While the LOD cloud
keeps growing, having a quick glimpse of those datasets is getting harder
and there is a need to develop new methods enabling to detect auto-
matically what an arbitrary dataset is about and to recommend visu-
alizations for data samples. We consider that “a visualization is worth
a million triples”, and in this paper, we propose a novel approach that
mines the content of datasets and automatically generates visualizations.
Our approach is directly based on the usage of SPARQL queries that will
detect the important categories of a dataset and that will specifically con-
sider the properties used by the objects which have been interlinked via
owl:sameAs links. We then propose to associate type of visualization for
those categories. We have implemented this approach into a so-called
Linked Data Vizualization Wizard (LDVizWiz).
Keywords: Visualization, human-machine Interaction, LOD, Consuming Linked
Data
1 Introduction
With the growing adoption of the Linked Data principles, there is a real need to
support data consumers in quickly getting visualizations that enable to explore a
dataset. In order to involve more general Web users into the Semantic Web and
Linked Data world, there is a need to build tools that reuse existing visualization
libraries showing the key information about RDF datasets. Many datasets are
published using SPARQL Endpoints and are not “visually” accessible. Thus,
understanding the underlying graphs and consuming them require lay users to
have some knowledge in writing queries.
The object of visualization is to develop insights from collected data. More-
over, according to Information Theory, vision is the sense that has the largest
bandwidth (100 Mbits/s), which makes it the best suited channel to convey in-
formation to the brain [12]. Based on the Visual Information Seeking Mantra:
“overview first, zoom and filter, then details on demand” [7], we advocate for more
2 Ghislain Auguste Atemezing, Raphaël Troncy
visual interactive representations of RDF graphs using SPARQL Endpoints. At
the same time, we use the term “Linked Data Visualization”, to refer to a com-
bination of charts, graphics, and other visual elements built on top of 4-5 stars
datasets accessible via a SPARQL endpoint. Despite the presence of more and
more datasets published as Linked Data, there is still a need to help end users to
discover what (unknown) datasets describe by hiding the complexity of SPARQL
queries from such users. Moreover, the task of identifying the key categories of
datasets can help in selecting and matching the most suitable visualization types.
In this paper, we propose a first step towards making available a semi-
automatic way for the production of possible visualization of linked data sets of
high-level categories grouping objects that are worth viewing and we associate
them with some very well known vocabularies. Then, we describe the implemen-
tation of a Linked Data Visualization Wizard and its main components. This
wizard can be used to easily visualize slices of datasets based on generic types
detected.
The remainder of this paper is structured as follows. We first provide some
related work that further motivates this study (Section 2). In Section 3, we
propose some important categories that are worth visualizing and we present
a set of mapping views associated with vocabularies (Section 4). In Section 5,
we describe the implementation of a wizard that can work on top of any RDF
dataset. We detail the results of an experiment where high level categories and
associated visualizations have been performed on numerous SPARQL endpoints
(Section 6). Finally, we conclude and present some future work in Section 7.
2 Related Work
There are currently many projects aiming at visualizing (RDF) Linked Data. A
survey by Dadzie and Rowe [2] concluded with the fact that many visualization
tools are not easy to use by lay users. In [4], there is a recent review of some
visualizations tools that can be summarized as follows:
– Vocabulary based visualization tools: these tools are built for specific vocab-
ularies and that help in visualizing data modelled according to those vocab-
ularies, such as CubeViz [6], FOAF explorer1 and Map4rdf [3]. They aim at
visualizing data modelled respectively with dq,foaf and geo+scovo.
– Mashup tools: they are used to create mashup visualizations with different
widgets and some data analysis, such as DERI Pipes [5]. Mashup tools can
be integrated into the LD wizard to combine different visual views.
– Generic RDF visualization tools: they typically support data browsing and
entity rendering. They can also be used to build applications. In this category,
we can mention Graphity2 , lodlive3 and Balloon Synopsis4 .
1
http://foaf-visualizer.gnu.org.ua/
2
https://github.com/Graphity/graphity-browser
3
http://en.lodlive.it/
4
https://github.com/schlegel/balloon-synopsis
Towards a Linked-Data based Visualization Wizard 3
While these tools are often extensible and support specific domain datasets, they
suffer from the following drawbacks:
– They are not easy to set up and use by lay users. Sometimes, users just need
to have a visual summary of a dataset in order to start exploring the data.
Our approach to this challenge is to provide such a lightweight javascript-
based tool that supports a quick exploration task.
– They do not make recommendation based on categories. A tool similar to
our approach is Facete5 [9] which shows a tree-based structure of a dataset
based on some properties of an endpoint more target at geodata. A tabular
view enables to visualize slices of data and a map view can be activated
when there is geo data. Our approach aims to be more generic, offering more
views (tabular, map, graph, charts, etc.) according to a systematic analysis
of what are the high level categories present in a dataset.
Regarding wizard-based tools for visualizing data, similar approaches are
available for tools consuming datasets in CSV/TSV. ManyEyes6 is an IBM online
tool that suggest charts according to the columns of a given CSV file. Similarly,
Google Charts7 help to achieve the same goals for creating embeddable charts
by using DSP language in the framework. Datawrapper8 is an open source tool
to enable the creation of basic charts originally target at journalists inspired
by ManyEyes and Google Charts. All the visualizations are based on the type
of the columns/fields of the data. In Linked Data, vocabularies are used for
modeling datasets in RDF, thus making it difficult to reuse directly those tools.
The benefit of our approach is that it constructs specific SPARQL queries to
detect the presence or not of predefined specific types of information, yielding
to information type-specific visualisations to enable end users to quickly start to
explore dataset in a generic manner.
3 Dataset Analysis
When developing an application, there are some “important” classes/categories,
objects or datatypes that can be detected first to help to guide in the progress of
creating a set of visualizations tied with those categories. We distinguish seven
categories while acknowledging that this is not necessary an exhaustive list:
– § [Geographic information]: This category is for data modeled using
geo:SpatialThing, dbpedia-owl:Place, schema:Place or gml:_Feature
classes.
– § [Temporal information]: This category also includes dataset containing
date, time (e.g: xsd:dateTime) and period or interval of time, using the
OWL Time ontology.
5
http://cstadler.aksw.org/facete/
6
http://www-958.ibm.com/software/data/cognos/manyeyes/
7
https://developers.google.com/chart/
8
http://datawrapper.de/
4 Ghislain Auguste Atemezing, Raphaël Troncy
– § [Event information]: This category is for any action of activity occurring
at some place at some time.
– § [Agent/Person information]: This category is heavily influenced by the
use of foaf:Person or foaf:Agent.
– § [Organization information]: This category is related to organizations or
companies data, with the use of the org vocabulary9 or the foaf:Organization
class.
– § [Statistics information]: This category refers to statistical data generally
modeled using the data cube vocabulary10 or the SDMX model11 .
– § [Knowledge Classification]: This category refers to dataset describing
schemas, classifications or taxonomies using the SKOS vocabulary.
4 Mapping Datatype, Views and Vocabularies
The On-line Library of Information Visualization Environments (OLIVE)12 is
a web site describing eight categories of information visualization environments
differentiated by data type and collected by students, following a visualization
course given at Maryland College Park, mostly inspired from the work of Ben
Shneiderman [7]. Based on the classification provided by OLIVE, we propose a
set of mappings between those categories (excluding the workspace dimension),
views that can be applied to this category and a suitable list of vocabularies
from the Linked Open Vocabularies catalogue [11]13 that correspond to those
categories. Those vocabularies are easy to be found as there is a manual classifi-
cation of vocabularies by the curators of the catalogue based on the content and
scope of the terms and properties. According to the seven categories defined in
Section 3, we have identified some of their corresponding one to one mapping
with the set of vocabularies:
– Geography space, consisting of 21 vocabularies for features: geo, gn, gf,
om, geop, md, lgdo, loc, igeo, osadm, geod, ostop, place, geos, locn, coun,
postcode, osr, geof, g50k and ad.
– Geometry space, for vocabularies dealing with the geometries, mostly com-
bined with the features, such as:
– Time space, consisting of 14 vocabularies, such as cal, date, gts, interval,
ncal, oh, te, thors, ti, time, tl, tm, tvc and tzont.
– Event space, containing vocabularies such as event, lode, music, sem, situ,
sport, stories, theatre, tis and tisc.
– Government space, with 9 vocabularies (cgov, ctorg, elec, few, gc, gd,
oan, odd, parl) and the org vocabulary belonging to the W3C recommen-
dation vocabularies at http://lov.okfn.org/dataset/lov/lov#W3C.
9
http://www.w3.org/TR/vocab-org/
10
http://www.w3.org/TR/vocab-data-cube/
11
http://sdmx.org/
12
http://lte-projects.umd.edu/Olive/
13
lov.okfn.org/dataset/lov/
Towards a Linked-Data based Visualization Wizard 5
Metadata vocabularies, such as rdfs, dcterms or dce can be used in associa-
tion with any of the visual element to give basic description of the resource of
a giving dimension. For example, a popup information can be fired on a map
view to display the relevant information of a geodata resource such as the label,
the abstract or description. Another application can be to detect which visual-
ization is best suited for geodata. Geodata belongs to a two-Dimension visual
representation. Geodata is usually displayed using geographical-based visualiza-
tions (map, geo charts, etc.) and it is often modeled by vocabularies in the space
named Geometry and Geography14 vocabularies in RDF datasets. Hence those
vocabularies can be combined to detect the presence or not of geographic infor-
mation in a dataset, and thus yield to recommend a map view. Table 1 gives an
overview of those mappings. For the tabular representation, it is the “default”
visual representation of RDF data and can be used by any vocabulary without
restriction.
Dimension Vocabulary Space Visual element
Temporal Time space TimeLine
one-Dimension any Tabular, text
two-Dimension Geography space Map view
Geometry space Maps view
three-Dimension Event space Map + TimeLine
Multi-Dimension qb, sdmx-model, scovo Charts, graphs
Tree skos, Government space Treemap, Org view
Network any vocab. Graph, network map
Table 1. A taxonomy of information visualization consuming Linked Datasets with
associated views and suitable vocabulary space.
5 LDVizWiz: a Linked Data Visualization Wizard
We propose a workflow composed of four principal steps for a Linked Data
Visualization Wizard, as depicted in Figure 1. Our requirement is to provide a
tool that hide the complexity of SPARQL to lay users and at the same time,
can be embedded in existing Linked Data infrastructure and workflow. First, we
proposed to detect the presence of data belonging to one of the seven categories
(Table 1) using generic SPARQL queries. More precisely, we perform ASK queries
to test whether or not a particular query pattern has a solution. Second, we look
at entities in a dataset that have owl:sameAs links with external objects and we
retrieve the properties associated to those objects. We argue that the objects that
are interlinked with other datasets are of primary importance in a visualization.
We show the results of this mining process to the user (the categories that
14
All the prefixes used for the vocabularies are the same used in LOV catalogue.
6 Ghislain Auguste Atemezing, Raphaël Troncy
have been detected, the properties going with the categories and the external
domain). Based on this information, the user can make a personalized “mashup”
by aggregating the suitable visualization widgets. Some default visualizations
are available according to the top categories detected. The last step is to publish
the visualization and a metadata report in RDF/XML TURTLE or N3.
Generate, publish and
Visualization share the visualizations
Publication Goal in interoperable formats
(HTML, SVG, RDF)
Mashup of different
Charts Goal charts based on the
Combination
properties and categories
Extract detected in the dataset
Detect Show
subject
SPARQL Endpoint
categories Sample data
properties Aggregation of properties
Property Goal used in interconnected
Aggregation datasets
Aggregate
Summarize
object + subject
information
properties Detection of main
Category
Detection Goal categories of the dataset
using ASK Queries.
Extract Extract object
sameAs links properties Visualize based on
categories detected SPARQL
endpoint
Detection Analysis Aggregation Publication
RDF RDF RDF
Dataset Dataset Dataset
(a) The workflow of the different modules (b) High level functionalities of the
interacting in the Linked Data visualization Linked Data visualization wizard
wizard.
Fig. 1. Big picture and architecture of the Linked Data visualization wizard.
Let consider a graph < G, c > to be G = {(s, p, o)|p ∈ U RI, s ∈ U RI, o ∈
(U RI ∪ LIT )} where U RI is the set of URIs, LIT is the set of literals, and c
the context. We define L = {V1 , V2 , ..., Vn |Vi = Pi ∪ Ti } the list of vocabularies
in LOV, with Pi and Ti respectively the properties and terms of a vocabulary
Vi . Let also D = {D1 , D2 , .., Dm } be the domains of vocabularies. We assume
∀V ∈ L, ∃Dk ∈ Φ(L, D). We define a generic function Σ : (G, c) 7→ B to detect
categories in a dataset as follows: Σ((G, c)) = {B|(∃(s, p, o) ∈ G : p ∈ V ) ∪
(∃(s, rdf : type, o) ∈ G : o ∈ V )} where B = {T rue, F alse}.
In the following sections, we describe each of the steps involved in the Linked
Data Visualization Wizard in more details.
5.1 Category Detection
The goal of the category detection task is to use SPARQL queries to detect the
presence of some high level categories in the dataset. We perform ASK queries as
implementation of the Σ function using standard vocabularies as defined in the
Table 1. We start with six categories, namely: geographic information, person,
Towards a Linked-Data based Visualization Wizard 7
organization, event, time and knowledge organization systems. We select popular
vocabularies based on two existing catalogues: LOV [10] and prefix.cc15 .
ASK WHERE {
{
?x a ?o .
f i l t e r ( ? o= dbpedia−owl : P l a c e | |
? o=gml : F e a t u r e | |
? o=g e o : S p a t i a l F e a t u r e | | ? o=gn : F e a t u r e | |
? o=admingeo : C i v i l A d m i n i s t r a t i v e A r e a | |
? o=s p a t i a l : F e a t u r e | |
? o=v c a r d : L o c a t i o n )
}
UNION {
? x ?p ? o . f i l t e r ( ? p=g e o : l a t | | ?p=g e o : l o n g | |
?p=g e o r s s : p o i n t | | ?p=g e o : geometry | |
geom : geometry )
}
}
Listing 1.1. Generic query to detect geo data from a SPARQL endpoint
Listing 1.1 shows seven classes of different vocabularies are used, respectively
for the namespaces dbpedia-owl, geo, gn, admingeo, spatial and vcard,
with relevant classes to check the presence of geographic data.
ASK WHERE {{? x a ? o . f i l t e r ( ? o=t i m e : T e m p o r a l E n t i t y | |
? o=t i m e : I n s t a n t | |
? o=t i m e : I n t e r v a l | | ? o=dbpedia−owl : TimePeriod | |
? o=t i m e : D a t e T i m e I n t e r v a l | | ? o=i n t e r v a l s : C a l e n d a r I n t e r v a l )
}
UNION{ ? x ?p ? o . f i l t e r ( ? p=t i m e : d u r a t i o n | |
?p=t i m e : h a s B e g i n n i n g | |
?p=t i m e : inDateTime | | ?p=t i m e : h a s D a t e T i m e D e s c r i p t i o n
| | ?p=t i m e : hasEnd ) }
Listing 1.2. Generic query to detect time data from a SPARQL endpoint, using time,
dbpedia-owl, intervals vocabularies.
Listing 1.2 detects the presence of time information, while Listing 1.3, 1.4
and 1.5 detect persons, organizations and events respectively.
ASK WHERE {? x a ? o . f i l t e r ( ? o = f o a f : P e r s o n ||
? o=dbpedia−owl : P e r s o n | |
? o=v c a r d : I n d i v i d u a l ) }
Listing 1.3. Generic query to detect person categories from a SPARQL endpoint,
using foaf, dbpedia-owl, vcard vocabularies.
PREFIX o r g :< h t t p : / /www. w3 . o r g / ns / o r g#>
PREFIX f o a f :
PREFIX dbpedia−owl :
ASK WHERE {? x a ? o . f i l t e r ( ? o=o r g : O r g a n i z a t i o n | |
? o=o r g : O r g a n i z a t i o n a l U n i t | |
? o=f o a f : O r g a n i z a t i o n | |
? o=dbpedia−owl : O r g a n i s a t i o n ) }
Listing 1.4. Generic query to detect ORG data from a SPARQL endpoint.
15
http://prefix.cc
8 Ghislain Auguste Atemezing, Raphaël Troncy
ASK WHERE{? x a ? o . f i l t e r ( ? o= l o d e : Event | | ? o=e v e n t : Event | |
? o=dbpedia−owl : Event ) }
Listing 1.5. Generic query to detect event data from a SPARQL endpoint, using lode,
event, dbpedia-owl vocabularies.
For detecting data organized as taxonomy, skos vocabulary is used along
with the most used classes and properties as showed in Listing 1.6.
ASK WHERE {{? x a ? o . f i l t e r ( ? o=s k o s : Concept | |
? o=s k o s : ConceptScheme | | ? o=s k o s : C o l l e c t i o n ) }
UNION{ ? x ?p ? o . f i l t e r ( ? p=s k o s : f e a t u r e C o d e | |
?p=s k o s : a l t L a b e l | | ?p=s k o s : p r e f L a b e l | | ?p=s k o s : r e l a t e d M a t c h ) } }
Listing 1.6. Generic query to detect SKOS data from a SPARQL endpoint, using skos
vocabulary.
5.2 Property Aggregation
We take the benefits of the owl:sameAs links between entities to have access
to the properties of the entities in the external namespaces different from the
origin dataset. This module also aggregates the properties found in the dataset
with the ones found in the interlinked sets. This is based on the assumption that
during the linkage process, external datasets not only help in not breaking the
follow-your-nose principle, but also add more information to be viewed in visual-
ization applications. As shown in the code below, at this stage, we have collected
and aggregated external properties gathered from the enrichment process of the
workflow.
1-LET Namespace(?s) = S and LET Namespace(?t) =T
2-SELECT owl:sameAs links
LET SEMTERM = list of ?s owl:sameAs ?t
WITH T != S
3-IN T, SELECT distinct properties used in dataset
4-AGGREGATE (3) with properties FROM S.
5.3 Visualization Generator
This module aims at recommending the appropriate visualizations based on the
categories detected by the wizard. It might also help the user to make a report
summarizing the result of the mining process, and then use different visualization
libraries to view the data. This module can be viewed as a recommender system
because it derives visualizations based on the categories. The input to build each
visualization is the corresponding SELECT query of each ASK queries used to
detect the categories. Moreover, some adjustment are made to avoid blank nodes
and to get the labels of the resources. The generator can be coupled with a
mashup widget generator for some categories. For example, users could expect
for event data, a combination of map view (where), a timeline (when) and facets
based on the agents (who).
Towards a Linked-Data based Visualization Wizard 9
5.4 Visualization Publisher
The publisher module aims at exporting the combined visualizations, along
with the report of all the process of mining the dataset, in a format easy to
share, either as HTML, SVG or in the different RDF syntax flavor. For the lat-
ter, apart from using metadata information (creator, issued date, license), we
model the categories we have detected using the dcterms:subject property of a
dcat:Dataset, the queries used (using the prov:wasDerivedFrom property), the
sample resources for each category (using the void:exampleResource property)
and the visualization generated (using the dvia and chart16 vocabularies).
6 Experiment and Implementation
In this section, we describe the experiments and report the evaluation on detect-
ing categories on 444 endpoints. We then describe a prototype as a“proof-of-concept”
of the proposal.
6.1 Experiment set up
We have evaluated our approach on the list of 444 endpoints referenced at
http://sparqles.okfn.org/ monitoring the availability, performance, inter-
operability and discoverability of SPARQL Endpoints registered in Datahub [1]
. We have implemented a script in python to speed up the process and obtain the
results. Every ASK query for the different category is implemented in a separate
function requesting a JSON response.
6.2 Evaluation
From the 444 endpoints used on the detection category module, 278 endpoints
(62.61%) were able to give satisfactory (yes/no on one of the seven categories)
results based on the queries. However, almost 37.38% of the endpoints were either
down at the time of our experiments or the response header was in XML instead
of JSON (as set up in the script). This result shows that our proposal with the
current implementation (not covering all the vocabularies in LOV) make use of
most popular vocabularies reused in the Linked Data.
This also implies a good coverage of the method that uses standard queries
and yet can be extended. The full result of the detection module on the queried
services is available at http://cf.datawrapper.de/3FuiV/2/, where for each
column, the value 0 stands for no presence and 1 for the presence of the cat-
egories. As provided in Table 2, 21.84% of geo data was detected, 13.288% of
person data, 10.81% of org data and 3.6% of SKOS data.
Table 3 summarizes some findings for 8 DBpedia chapters endpoints where
it’s easy to note the absence of SKOS data, and less presence of data modeled
using time vocabulary. The Table also shows the differences in the standard vo-
cabularies used to convert the wikipedia data into RDF across different chapters.
16
http://data.lirmm.fr/ontologies/chart
10 Ghislain Auguste Atemezing, Raphaël Troncy
Category number Percentage
GEO DATA 97 21.84%
EVENT DATA 16 3.60%
TIME DATA 27 6.08%
SKOS DATA 2 0.45%
ORG DATA 48 10.81%
PERSON DATA 59 13.28%
STAT DATA 29 6.6%
Table 2. Classification of the endpoints according to the datatype detected with our
SPARQL generic queries
Endpoint event geo org person skos time
dbpedia.org 0 1 1 1 0 0
de.dbpedia.org 0 1 1 1 0 0
el.dbpedia.org 1 1 1 1 0 0
fr.dbpedia.org 1 1 1 1 0 1
ja.dbpedia.org 1 1 1 1 0 0
live.dbpedia.org 1 1 1 1 0 1
nl.dbpedia.org 1 1 1 1 0 0
pt.dbepdia.org 1 1 1 1 0 0
Table 3. Categories detected in some dbpedia endpoint domains, where “1” is the
presence and “0” the absence of the given type of category.
6.3 Implementation
A first prototype, implemented with javascript and the Bootstrap framework17 ,
is available at http://semantics.eurecom.fr/datalift/rdfViz/apps/, as a
proof of concept. We aim at providing a lightweight tool for lay users to quickly
understand what the data is about and so that they get first visualizations based
on categories detected in the datasets. We also reuse sgvizler [8] for generating
charts according to the categories retrieved by the wizard. In the current imple-
mentation, the user can enter any SPARQL endpoint, and with a “click”, the user
can receive the list of categories detected together with sample resources. In the
second step, the wizard retrieves the properties from the objects and subjects
part of owl:sameAs links. The last step shows different tabs with the summary
of the previous steps, the visualizations available for each categories, and a re-
port both in human and machine readable formats. Figure 2 depicts a sample
visualization generated by the wizard for geo data and statistics data.
The system can be used in any tool consuming Linked Data in which the
complexity of SPARQL analysis and visualizations of RDF datasets is hidden to
the lay users, with the benefits of showing that information encoded in triples
is not only ”beautiful”, but also useful in the sense of traditional wizard-based
tools.
17
http://getbootstrap.com/
Towards a Linked-Data based Visualization Wizard 11
(a) Overview of the first step report (b) Two views generated by the LD
on categories detected:(1) categories wizard, for geodata and stats data
detected,(2) vocabularies used.
Fig. 2. Categories detected and visualization generated by the Linked Data visualiza-
tion wizard in the case of EventMedia endpoint service.
7 Conclusion and Future Work
In this paper we have presented an approach for creating visualizations on top
of Linked Data. We first defined seven categories of objects worth viewing in a
dataset, and we propose to associate them with common Linked Data vocabu-
laries. We then present a description of the main components of LDVizWiz: a
Linked Data Visualization Wizard. We describe a lightweight implementation in
javascript as a proof-of-concept of our proposal, with the benefits to be usable
on-line or being extensible. We advocate that such a tool can be easily integrated
in any workflow for publishing and linking data on the web, like Datalift18 or the
GeoKnow stack19 . We plan to use a more exhaustive set of vocabularies in our
generic queries for detecting those categories, plugging directly the wizard to the
LOV catalogue. Regarding the aggregation properties, it can be extended to take
into account other semantic relations (e.g: skos:exactMatch). Additionally, we
plan to make an evaluation of the prototype and compare it to related tools such
as the ones aiming to build a dataset profile.
8 Acknowledgments
This work has been partially supported by the French National Research Agency
(ANR) within the Datalift Project, under grant number ANR-10-CORD-009.
18
http://www.datalift.org
19
http://generator.geoknow.eu/
Bibliography
[1] C. Buil-Aranda, A. Hogan, J. Umbrich, and P.-Y. Vandenbussche. Sparql
web-querying infrastructure: Ready for action? In The Semantic Web–
ISWC 2013, pages 277–293. Springer, 2013.
[2] A.-S. Dadzie and M. Rowe. Approaches to visualising linked data: A survey.
Semantic Web Journal, 2(2):89–124, 2011.
[3] A. de Leon, F. Wisniewki, B. Villazón-Terrazas, and O. Corcho. Map4rdf
- Faceted Browser for Geospatial Datasets. In Using Open Data: policy
modeling, citizen empowerment, data journalism (PMOD’12), 2012.
[4] J. Klimek, J. Helmich, and M. Neasky. Application of the Linked Data
Visualization Model on Real World Data from the Czech LOD Cloud. In
6th International Workshop on the Linked Data on the Web (LDOW’14),
2014.
[5] D. L. Phuoc, A. Polleres, C. Morbidoni, M. Hauswirth, and G. Tummarello.
Rapid semantic web mashup development through semantic web pipes.
In 18th International World Wide Web Conference (WWW’09), Madrid,
Spain, 2009.
[6] P. E. Salas, M. Martin, F. M. D. Mota, K. Breitman, S. Auer, and M. A.
Casanova. Publishing statistical data on the web. In Proceedings of 6th
International IEEE Conference on Semantic Computing, IEEE 2012. IEEE,
2012.
[7] B. Shneiderman. The Eyes Have It: A Task by Data Type Taxonomy for
Information Visualizations. In IEEE Symposium on Visual Languages ’96,
IEEE, Los Alamos, CA (September 1996), pages 336–343, 1996.
[8] G. M. Skjaeveland. Sgvizler: A javascript wrapper for easy visualization of
sparql result sets. In 9th Extended Semantic Web Conference (ESWC’12),
2012.
[9] C. Stadler, M. Martin, and S. Auer. Exploring the Web of Spatial Data
with Facete. In Companion proceedings of 23rd International World Wide
Web Conference (WWW), pages 175–178, 2014.
[10] P.-Y. Vandenbussche and B. Vatant. Metadata Recommendations For
Linked Open Vocabularies. OKFN, 2012. http://lov.okfn.org/dataset/
lov/Recommendations_Vocabulary_Design.pdf.
[11] P.-Y. Vandenbussche, B. Vatant, and L. Rozat. Linked open vocabularies:
an initiative for the web of data. In QetR Workshop, Chambery, France,
2011.
[12] C. Ware. Information Visualization, Second Edition: Perception for Design.
Morgan Kaufmann Publishers Inc.; 2 edition (April 21, 2004), San Francisco,
CA, USA, 2014.