Conference PaperPDF Available

The Linked Data Visualization Model

Authors:

Figures

Content may be subject to copyright.
The Linked Data Visualization Model
Josep Maria Brunetti1, Sören Auer2, Roberto García1
1GRIHO, Universitat de Lleida
Jaume II, 69. 25001 Lleida, Spain
{josepmbrunetti,rgarcia}@diei.udl.cat,http://griho.udl.cat/
2AKSW, Computer Science
University of Leipzig, Germany
auer@informatik.uni-leipzig.de,http://aksw.org/
Abstract. In the last years, the amount of semantic data available in
the Web has increased dramatically. The potential of this vast amount of
data is enormous but in most cases it is very difficult for users to explore
and use this data, especially for those without experience with Semantic
Web technologies. Applying information visualization techniques to the
Semantic Web helps users to easily explore large amounts of data and
interact with them. The main objective of information visualization is to
transform and present data into a visual representation, in such a way
that users can obtain a better understanding of the data. However, the
large amount of Linked Data available and its heterogeneity makes it
difficult to find the right visualization for a given dataset. In this article
we devise a formal Linked Data Visualization model (LDVM), which al-
lows to dynamically connect data with visualizations. In order to achieve
such flexibility and a high degree of automation the LDVM is based on
a visualization workflow incorporating analytical extraction and visual
abstraction steps. We report about our comprehensive implementation
of the Linked Data Visualization model comprising a library of generic
visualizations that enable to get an overview on, visualize and explore
the Data Web.
Keywords: Semantic Web, Linked Data, Visualization, Interaction
1 Introduction
In the last years, the amount of semantic data available in the Web has increased
dramatically, especially thanks to initiatives like Linked Open Data (LOD). The
potential of this vast amount of data is enormous but in most cases it is very dif-
ficult and cumbersome for users to visualize, explore and use this data, especially
for lay-users [1] without experience with Semantic Web technologies.
Visualizing and interacting with Linked Data is an issue that has been rec-
ognized from the beginning of the Semantic Web (cf. e.g. [2]). Applying informa-
tion visualization techniques to the Semantic Web helps users to explore large
amounts of data and interact with them. The main objectives of information
visualization are to transform and present data into a visual representation, in
such a way that users can obtain a better understanding of the data [3]. Visu-
alizations are useful for obtaining an overview of the datasets, their main types,
properties and the relationships between them.
Compared to prior information visualization strategies, we have a unique
opportunity on the Data Web. The unified RDF data model being prevalent
on the Data Web enables us to bind data to visualizations in an unforeseen
and dynamic way. An information visualization technique requires certain data
structures to be present. When we can derive and generate these data structures
automatically from reused vocabularies or semantic representations, we are able
to realize a largely automatic visualization workflow. Ultimately, we aim to re-
alize an ecosystem of data extractions and visualizations, which can be bound
together in a dynamic and unforeseen way. This will enable users to explore
datasets even if the publisher of the data does not provide any exploration or
visualization means.
Most of existing work related with visualizing RDF is focused on concrete
domains and concrete datatypes. However, finding the right visualization for a
given dataset is not an easy task [4]: ‘One must determine which questions to
ask, identify the appropriate data, and select effective visual encodings to map
data values to graphical features such as position, size, shape, and color. The
challenge is that for any given data set the number of visual encodings – and
thus the space of possible visualization designs – is extremely large.’
The Linked Data Visualization Model we propose in this paper allows to
connect different datasets with different visualizations in a dynamic way. In
order to achieve such flexibility and a high degree of automation the LDVM is
based on a visualization workflow incorporating analytical extraction and visual
abstraction steps. Each of the visualization workflow steps comprises a number of
transformation operators, which can be defined in a declarative way. As a result,
the LDVM balances between flexibility of visualization options and minimality
of configuration.
Our main contributions are in particular:
1. The adoption of the Data State Reference Model [5] for the RDF data model.
2. The creation of a formal Linked Data Visualization model, which allows to
dynamically connect data with visualizations.
3. A comprehensive implementation of the Linked Data Visualization model
comprising a library of generic visualizations that enable to get an overview
on, visualize and explore the Data Web.
The remainder of this paper is structured as follows: Section 2 discusses re-
lated work. Section 3 introduces the Linked Data Visualization Model. Section 4
describes an implementation of the model and Section 5 presents its evaluation
with different datasets and visualizations. Finally, Section 6 contains conclusions
and future work.
2 Related Work
Related work can be roughly classified into tools supporting Linked Data visu-
alization and exploration, data visualization in general as well as approaches for
visualizing and interacting with particularly large datasets.
2.1 Linked Data Visualization and Exploration
Exploring and visualizing Linked Data is a problem that has been addressed
by several projects. Dadzie et al. [1] analyzed some of those projects concluding
that most of them are designed only for tech-users and do not provide overviews
on the data.
Linked Data browsers such as Disco [6], Tabulator [7] or Explorator [8] al-
low users to navigate the graph structures and usually display property-value
pairs in tables. They provide a view of a subject, or a set of subjects and their
properties, but not any additional support getting a broader view of the dataset
being explored. Rhizomer [9] provides an overview of the datasets and allows to
interact with data through Information Architecture components such as navi-
gation menus, breadcrumbs and facets. However, it needs to precompute some
aggregated values and does not include visualizations.
Graph-based tools such as Fenfire [10], RDF-Gravity1,IsaViz 2provide node-
link visualizations of the datasets and the relationships between them. Although
this approach can help obtaining a better understanding of the data structure,
in some cases graph visualization does not scale well to large datasets [11]. Some-
times the result is a complex graph difficult to manage and understand [12].
There are also JavaScript-based tools to visualize RDF. Sgvizler 3renders
the results of SPARQL queries into HTML visualizations such as charts, maps,
treemaps, etc. However, it requires SPARQL knowledge in order to create RDF
visualizations. LODWheel [13] also provides RDF visualizations for different data
categories but it is focused on visualizing vocabularies used in DBpedia.
Other tools are restricted to visualizing and browsing concrete domains, e.g.
LinkedGeoData browser [14] or map4rdf 4for spatial data, FoaF Explorer5to
visualize RDF documents adhering to the FOAF vocabulary.
To summarize, most of the existing tools make it difficult for non-technical
users to explore linked data or they are restricted to concrete domains. None
of them provide generic visualizations for RDF data. Consequently, it is still
difficult for end users to obtain an overview of datasets, comprehend what kind
of structures and resources are available and what properties resources typically
have and how they are mostly related with each other.
1http://semweb.salzburgresearch.at/apps/rdf-gravity
2http://www.w3.org/2001/11/IsaViz
3http://code.google.com/p/sgvizler/
4http://oegdev.dia.fi.upm.es/map4rdf/
5http://xml.mfd-consult.dk/foaf/explorer/
2.2 Data Visualization
Regarding data visualization, most of the existing work is focused on categorizing
visualization techniques and systems [15]. For example, [16] and [17] propose
different taxonomies of visualization techniques. However, most visualization
techniques are only compatible with concrete domains or data types [3].
Chi’s Data State Reference Model [5] defines the visualization process in a
generic way. It describes a process for transforming raw data into a concrete vi-
sualization by defining four data stages as well as a set of data transformations
and operators. This framework provides a conceptual model that allows to iden-
tify all the components in the visualization process. In Section 3 we describe how
this model serves as a starting point for our Linked Data Visualization Model.
2.3 Visualizing and interacting with large-scale datasets
For data analysis and information visualization, Shneiderman proposed a set
of tasks based on the visual information seeking mantra: “overview first, zoom
and filter, then details on demand [18]. Gaining overview is one of the most
important user tasks when browsing a dataset, since it helps end users to un-
derstand the overall data structure. It can also be used as starting point for
navigation. However, overviews become difficult to achieve with large heteroge-
neous datasets, which is typical for Linked Data. A common approach to obtain
an overview and support the exploration of large datasets is to structure them
hierarchically [19]. Hierarchies allow users to visualize different abstractions of
the underlying data at different levels of detail. Visual representations of hier-
archies allow to create simplified versions of the data while still maintaining the
general overview.
There are several techniques for visualizing hierarchical structures. One ap-
proach to provide high level overviews are Treemaps [20]. Treemaps use a rect-
angle to show the tree root and its children. Each child has a size proportional
to the cumulative size of its descendants. They are a good method to display
the size of each node in a hierarchy. In CropCircles [21] each node in the tree is
represented by a circle. Every child circle is nested inside its parent circle. The
diameter of the circles are proportional to the size of their descendants. Space-
Trees [22] combines the conventional layout of trees with zooming interaction.
Branches of the tree are dynamically rescaled to fit the available space.
For datasets containing spatial data, another approach for obtaining an
overview is to point resources on a map. Clustering [23] and organizing resources
into hierarchies [24] are the most commonly used approaches for exploring large
datasets on maps.
By applying these techniques it is possible to build interactive hierarchical
visualizations that support the visual information seeking mantra proposed by
Shneiderman.
3 Linked Data Visualization Model
In this section we present the Linked Data Visualization Model by first giving
an overview, then formalizing key elements of the model and finally showcasing
the model in action with a concrete example.
3.1 Overview
We use the Data State Reference Model (DSRM) proposed by Chi [5] as con-
ceptual framework for our Linked Data Visualization Model (LDVM). While the
DSRM describes the visualization process in a generic way, we instantiate and
adopt this model with LDVM for the visualization of RDF and Linked Data.
The names of the stages, transformations and operators have been adapted to
the context of Linked Data and RDF. Figure 1 shows an overview of LDVM. It
that can be seen as a pipeline, which originates in one end with raw data and
results in the other end with the visualization.
RDF DATA
ANALYTICAL EXTRACTION
VISUALIZATION ABSTRACTION
VIEW
DATA
TRANSFORMATION
VISUALIZATION
TRANSFORMATION
VISUAL MAPPING
TRANSFORMATION
SPARQL
OPERATORS
ANALYTICAL
OPERATORS
VISUALIZATION
OPERATORS
VIEW
OPERATORS
Fig. 1. High level overview of the Linked Data Visualization Model.
The LDVM pipeline is organized in four stages that data needs to pass
through:
1. RDF Data: the raw data, which can be all kinds of information adhering to
the RDF data model, e.g. instance data, taxonomies, vocabularies, ontolo-
gies, etc.
2. Analytical extraction: data extractions obtained from raw data, e.g. calcu-
lating aggregated values.
3. Visual abstraction: information that is visualizable on the screen using a
visualization technique.
4. View: the result of the process presented to the user, e.g. plot, treemap, map,
timeline, etc.
Data is propagated through the pipeline from one stage to another by apply-
ing three types of data transformation operators:
1. Data transformation: transforms raw data values into analytical extractions
declaratively (using SPARQL query templates).
2. Visualization transformation: takes analytical extractions and transforms
them into a visualization abstraction. The goal of this transformation is to
condense the data into a displayable size and create a suitable data structure
for particular visualizations.
3. Visual mapping transformation: processes the visualization abstractions in
order to obtain a visual representation.
There are also operators within each stage that do not change the underlying
data but allow to extract data from each stage or add new information. These
are:
1. SPARQL Operators: SPARQL functions, e.g. SELECT, FILTER, COUNT,
GROUP BY, etc.
2. Analytical Operators: functions applied to the data extractions, e.g. select a
subset of data.
3. Visualization Operators: visual variables, e.g. visualization technique, sizes,
colors, etc.
4. View Operators: actions applied to the view, e.g. rotate, scale, zoom, etc.
As shown in Figure 2, our model allows to connect different RDF datasets and
different data extractions with different visualization techniques. Not all datasets
are compatible with all data extractions, as well as each data extraction is only
compatible with some visual configurations.
Each dataset has different data extraction possibilities, e.g. class hierarchy,
property hierarchy, geospatial data, etc. Each data extraction can be visualized
with different configurations, which contain information such as the visualiza-
tion technique to use, colors, etc. Then, a concrete visualization is generated
depending on the data extraction and the visual configuration.
In summary, the model is divided in two main blocks: data space and visual
space. The RDF data stage, analytical extraction stage and data transformation
belong to the data space, while visual abstraction stage, view stage and visual
mapping transformation belong to the visual space. These two main blocks are
connected by a visualization transformation.
Dataset 1
Dataset 2
Dataset n
Data 1
Data 2
Data n
Config 1
Config 2
Config n
.
.
.
View 1
View 2
View n
.
.
.
.
.
.
.
.
.
Datasets Data Configuration View
IncompatibleCompatible
RDF Data Analytical Extraction Visualization Abstraction View
Data Transformation Visualization Transformation Visual Mapping Transformation
Fig. 2. Linked Data Visualization Model workflow.
3.2 Formalization
The RDF data model enables us to bind data to visualizations in a dynamic
way while the LDVM allows to connect datasets with data extractions and visu-
alizations. However, the compatibility between data and concrete visualizations
needs to be determined.
In this section we formalize the extraction and generation of visualization
data structures (in the Analytical Extraction Stage) as well as the visualization
configuration (in the Visualization Abstraction Stage). Our visual data extrac-
tion structure is capturing the information required to (a) determine whether a
certain dataset is compatible with a certain analytical extraction and (b) process
a compatible dataset and extract a relevant subset (or aggregation) of the data
as required by compatible visualization configurations. Our visual configuration
structure determines wether a certain analytical extraction is compatible with a
certain visualization technique.
Definition 1 (Visualization Data Extraction). The Visualization Data Ex-
traction defines how data is accessed for the purpose of visualizations. Our Vi-
sualization Data Extraction structure comprises the following elements:
a compatibility check function (usually a SPARQL query), which determines
whether a certain dataset contains relevant information for this data extrac-
tion,
a possibly empty set Dof data extraction parameters,
a set tuples (q, σq), with qbeing a SPARQL query template containing place-
holders mapped to the data extractions parameters Dand σqbeing the sig-
nature of the SPARQL query result, i.e. the number of columns returned by
queries generated from qand their types.
Definition 2 (Visualization Configuration). The Visualization Configura-
tion is a data structure, which accommodates all configuration information re-
quired for a particular type of visualization. This includes:
the visualization technique, e.g. Treemap, CropCircles, etc,
visualization parameters, e.g. colors, aggregation,
an aggregated data assembly structure, consisting of a set of relations r, each
one equipped with a signature σr
Definition 3 (Compatibility). A Visualization Data Extraction vde is said
to be compatible with a Visualization Configuration vc iff for each σrin vc there
is a corresponding σqin vde, such that the types and their positions in σqand
σrmatch.
3.3 Example application of the model
In this section we present a concrete application of the Linked Data Visualization
Model. It describes the process to create a Treemap visualization of the DBpedia
[25] class hierarchy with its Data Stages, Transformations and Operators used
in each Data Stage.
Data Stages
RDF Data: the DBpedia dataset containing RDF triples and its ontology.
Analytical Extraction: tuples containing the class URI, its number of in-
stances and its children.
Visual Abstraction: data structure containing the class hierarchy.
View: a treemap representation of the class hierarchy.
Transformations
Data Transformation: SPARQL queries to obtain the analytical extraction.
The concrete SPARQL queries can be seen in the Example 1.
Visual Transformation: a function that creates the class hierarchy from the
tuples obtained in the Analytical Extraction Stage.
Visual Mapping Transformation: a function that transforms the class hier-
archy into a Treemap using a concrete visualization library.
Stage Operators
SPARQL Operators: DISTINCT, COUNT, GROUP BY, AS, FILTER, OP-
TIONAL.
Analytical Extraction Operators: functions to parse the SPARQL queries
results and convert them to a concrete format and structure.
Visual Abstraction Operators: size of the treemap, colors.
View: zoom in, zoom out.
Example 1 (Visualization Data Extraction).
Compatibility check function: the following SPARQL query is executed in
order to check if this data extraction is compatible with the dataset. If the query
returns a result it means that this information is present in the dataset and the
data extraction is compatible with it.
1AS K
2WH E R E {
3? c rd f : ty p e ?c l as s .
4FI L TE R (? c la ss = o w l: C l as s || ? c la ss = r d fs : C la s s )
5}
6LI M I T 1
SPARQL Query to obtain root elements: the following SPARQL query ob-
tains the root elements of the class hierarchy and their labels if they exist.
1SE L EC T D I ST I NC T ? ro o t ?l a be l
2WH E R E {
3? ro o t rd f : ty p e ?c l as s .
4FI L TE R (? c la ss = o w l: C l as s || ? c la ss = r d fs : C la s s ) .
5OP T I ON A L {? r oo t r df s : s ub C la s s Of ? s up e r .
6FI L TE R (? r oo t != ? s up e r && ? su p er ! = ow l : T hi ng & & ?s u pe r != r df s :
Re s ou r ce & & ! is Bl a nk ( ? su p er ) )
7} .
8OP T I O N A L {
9? ro o t rd f s : la b el ? l ab e l .
10 FI L TE R ( L AN G (? l ab el ) = ’ en ’ || L AN G ( ? la b el ) = ’ ’)
11 } .
12 FI L TE R (! b ou nd ( ? su pe r ) && is UR I (? r oo t ) && ! is Bl an k (? r o ot ) && ? ro ot
!= o w l : Th i ng )
13 }
SPARQL Query to obtain children per element: the following SPARQL query
is used to obtain the parent of each class in the hierarchy and their labels if they
exist.
1SE L E CT D I ST I N CT ? s ub ? l ab e l ?p a re n t
2WH E R E {
3? su b r df : t y pe ? c la s s .
4FI L TE R (? c la ss = o w l: C l as s || ? c la ss = r d fs : C la s s ) .
5? su b r df s : s ub C l as s Of ? p a re n t .
6FI L TE R (! i sB la n k (? p ar en t ) && ?p a re nt ! = ow l: T hi ng ) .
7OP T I O N A L {
8? su b r df s : s ub C la s sO f ? s ub 2 .
9? su b 2 rd f s : su b Cl a s sO f ? p ar e nt .
10 OP T I O N A L {
11 ? p ar e nt o w l : eq u i va l e nt C l as s ? s ub 3 .
12 ? su b 3 rd f s : su b Cl a s sO f ? su b 2
13 } .
14 FI L T E R ( ? s ub ! =? s ub2 & & ? s u b2 ! =? p aren t && ! i s Bla n k (? su b 2 ) & & !
bo u nd (? s ub 3 ) )
15 } .
16 OP T I O N A L {
17 ? su b r df s : l ab e l ? la b el .
18 FI L TE R ( L AN G (? l ab el ) = ’ en ’ )
19 } .
20 FI L TE R (! i sB la nk ( ? su b ) & & !b ou nd ( ? su b2 ) )
21 }
SPARQL Query to obtain number of resources per element: the following
SPARQL query obtains the number of resources of each class in the hierarchy.
1SE L E C T ? c ( C O UNT (? x) AS ? n )
2WH E R E {
3? x rd f : ty p e ?c .
4? c rd f : ty p e ?c l as s .
5FI L TE R (? c la ss = o w l: C l as s || ? c la ss = r d fs : C la s s )
6}
7GR O U P B Y ? c
SPARQL Query to obtain a list of resources per element: the following SPARQL
query obtains a list of resources that belong to a concrete class (in placeholder
%s)
1SE L EC T D IS T IN C T ?x ? la b el
2WH E R E {
3? x rd f : ty p e %s .
4OP T I O N A L {
5? x rd fs : l ab e l ? la be l .
6FI L TE R ( L AN G (? l ab el ) = ’ en ’ || L AN G ( ? la b el ) = ’ ’)
7}
8}
9LI M I T 10
Example 2 (Visualization Configuration). The visualization Configuration struc-
ture contains:
Visualization Technique: Treemap,
Visualization parameters:
Colors: random,
Aggregation6: yes.
4 Implementation
Based on the Linked Data Visualization Model proposed, we have implemented
a prototype called LODVisualization7. It allows to explore and interact with
the Data Web through different visualizations. The visualizations implemented
suport the Overview task proposed by Shneiderman [18] in his visual information
seeking mantra. This way, our prototype serves not only as a proof of concept of
our LDVM but also provides useful visualizations of RDF. These visualizations
allow users to obtain an overview of RDF datasets and realize what the data is
about: their main types, properties, etc.
6group those classes that have an area too small to show in the Treemap
7http://lodvisualization.appspot.com/
LODVisualization is developed using Google App Engine8(GAE). Google
App Engine is a cloud computing platform for developing and hosting web ap-
plications in Google’s infrastructure. Figure 3 shows the architecture of LODVi-
sualization.
SPARQL
Endpoint
PROXY
SPARQL
Query Cache
SPARQL
Endpoint
SPARQL
Query
SPARQL
Query
JSON
JSON
BACKEND
Google App Engine (Python)
FRONTEND
HTML + CSS + JAVASCRIPT
GAE
Datastore
GAE
Blobstore
1
2
3
4
5
Graph
Selection
Data
Selection
Visualization
Selection
Results
Fig. 3. LODVisualiation architecture
Steps 1 and 2 correspond to the RDF Data Stage, in which users can enter an
SPARQL endpoint and select those graphs to visualize. Step 3 corresponds to the
Data Extraction Stage, in which users can choose a data extraction available for
the graphs selected. Step 4 corresponds to the Visualization Abstraction Stage,
in which users can select a visualization technique compatible with the data
extraction and some visualization parameters such as colors, size, etc. Finally,
step 5 corresponds to the View Stage and presents a concrete visualization to
the users.
The frontend is developed using HTML, CSS and Javascript, while the back-
end is developed in Python. Most visualizations are created using JavaScript
Infovis Toolkit9and D3.js [26]. We have chosen them because they allow to cre-
ate rich and interactive JavaScript visualizations and they include also a free
license. We also use Google Maps and OpenStreetMap to create Geospatial Vi-
sualizations. Data is transferred between SPARQL endpoints, the server and the
8https://developers.google.com/appengine/
9http://thejit.org/
client using JSON, which is the data format used by those libraries. Most of
SPARQL endpoints support JSON and can convert RDF to it, making it easy
to integrate it into Javascript.
Instead of querying directly SPARQL endpoints using AJAX, queries are
generated in the client side but sent to the server. The server acts as proxy in
order to avoid AJAX Cross-domain requests. At the same time, the server pro-
vides a cache memory implemented using GAE Datastore and GAE Blobstore.
For each SPARQL query we generate a MD5 hash and we store it with the query
results. The Blobstore is used to save those results whose size exceeds 1Mb be-
cause the Datastore has this limitation. When the server receives a query, the
results are obtained from cache if it is already there. Otherwise, the query is
sent to the SPARQL endpoint using the GAE URL Fetch API. Then, the hash
of the query and its results are stored in cache. A cache memory is useful to
prevent executing the same queries when the users want to visualize the same
data extraction with different visualization techniques.
Our implementation includes so far 4 different data extraction types and
7 different visualization techniques. The data extraction possibilities are: class
hierarchy, property hierarchy, SKOS concepts hierarchy and geospatial data.
The visualization techniques implemented are: Treemap, Spacetree, CropCircles,
Indented Tree Google Maps and OpenStreetMap. Figure 4 shows a concrete
visualization example with the following parameters:
Dataset: DBpedia.
Data Extraction: class hierarchy.
Visualization Configuration: Treemap, aggregation, random colors.
Fig. 4. DBpedia class hierarchy Treemap
LODVisualization is compatible with most of SPARQL endpoints as long
as they support JSON and SPARQL 1.110. Most of the data extraction queries
use aggregate functions such as COUNT or GROUP BY, which are implemented
from this version. Our implementation has a limitation: the response of each
SPARQL query must be in less than 60 seconds. This is the maximum deadline
for requests using the GAE URL Fetch API. Nevertheless, longer requests could
be performed using the GAE Task Queue API.
5 Evaluation
We have evaluated our implementation of the Linked Data Visualization Model
with different datasets, data extractions and visualizations. The evaluation was
performed with Google Chrome version 19.0.1084.46. Table 1 shows a summary
of the evaluation results.
Dataset
Data Extraction
Visualization
Configuration
Execution time (s)
w/o cache
w/ cache
DBpedia
Class hierarchy
Treemap
3.58
0.87
DBpedia
Class hierarchy
Crop Circles
3.35
0.76
DBpedia
SKOS concept hierarchy
Treemap
48.79
16.96
DBpedia
Spatial data
Google Maps
2.05
0.37
Dbpedia
Spatial data
OpenStreetMaps
2.03
0.79
Wine Ontology
Property hierarchy
Indented Tree
1.95
0.57
Wine Ontology
Property hierarchy
Space Tree
2.45
0.70
LinkedMDB
Class hierarchy
Treemap
3.28
0.61
WWW 2011
Class hierarchy
Crop Circles
2.53
0.75
AGRIS FAO
SKOS concept hierarchy
Treemap
28.72
8.54
Data Transformation -
SPARQL Query Templates (s)
Visual
Transformation (s)
Visual Mapping
Transformation (s)
#1
#2
#3
Total
1.4336
1.2741
0.7305
3.4382
0.0190
0.1228
1.2194
1.3423
0.7123
3.2740
0.0140
0.0620
3.5202
5.5424
35.9858
45.0484
0.4588
3.2900
2.0028
-
-
2.0028
0.0148
0.0356
1.6008
-
-
1.6008
0.0100
0.4266
0.7792
0.4860
0.6712
1.9364
0.0022
0.0130
0.98875
0.6757
0.5650
2.2295
0.0015
0.0372
0.9554
0.7334
1.5442
3.233
0.0086
0.0482
0.8914
0.8924
0.7016
2.4854
0.0102
0.0430
6.2942
6.4924
14.1782
26.9648
0.5428
1.2184
Table 1. Evaluation results summary
Each concrete visualization was executed 10 times without cache, 10 times
with cache and we calculated the average time of the executions. As it can be
seen, most of the visualizations can be generated in real time and the use of a
cache memory reduces the execution time. However, in experiments #3 and #10
(SKOS concept hierarchies) the results had to be stored in the GAE Blobstore
instead of the GAE Datastore due to their size. Accessing the Blobstore to
retrieve those results is much slower than accessing the Datastore.
Table 2 shows the timings for each visualization divided into the three trans-
formations proposed in the LDVM: data transformation, visual transformation
and visual mapping transformation.
10 http://www.w3.org/TR/sparql11-query/
ID
Dataset
Data Extraction
Visualization
Configuration
Execution time (s)
w/o cache
w/ cache
1
DBpedia
Class hierarchy
Treemap
3.58
0.87
2
DBpedia
Class hierarchy
Crop Circles
3.35
0.76
3
DBpedia
SKOS concept hierarchy
Treemap
48.79
16.96
4
DBpedia
Spatial data
Google Maps
2.05
0.37
5
Dbpedia
Spatial data
OpenStreetMaps
2.03
0.79
6
Wine Ontology
Property hierarchy
Indented Tree
1.95
0.57
7
Wine Ontology
Property hierarchy
Space Tree
2.45
0.70
8
LinkedMDB
Class hierarchy
Treemap
3.28
0.61
9
WWW 2011
Class hierarchy
Crop Circles
2.53
0.75
10
AGRIS FAO
SKOS concept hierarchy
Treemap
28.72
8.54
ID
Data Transformation -
SPARQL Query Templates (s)
Visual
Transformation (s)
Visual Mapping
Transformation (s)
#1
#2
#3
Total
1
1.4336
1.2741
0.7305
3.4382
0.0190
0.1228
2
1.2194
1.3423
0.7123
3.2740
0.0140
0.0620
3
3.5202
5.5424
35.9858
45.0484
0.4588
3.2900
4
2.0028
-
-
2.0028
0.0148
0.0356
5
1.6008
-
-
1.6008
0.0100
0.4266
6
0.7792
0.4860
0.6712
1.9364
0.0022
0.0130
7
0.98875
0.6757
0.5650
2.2295
0.0015
0.0372
8
0.9554
0.7334
1.5442
3.233
0.0086
0.0482
9
0.8914
0.8924
0.7016
2.4854
0.0102
0.0430
10
6.2942
6.4924
14.1782
26.9648
0.5428
1.2184
Table 2. Timing for each transformation
Creating different visualizations for the same data extraction takes a similar
time. This is because most of the execution time corresponds to the data trans-
formation. The visual transformation timing depends on the size of the results
to process. In the same way, the visual mapping transformation depends on the
number of items to visualize. This is particularly high in experiments #3 and
#10, with a huge hierarchy of SKOS concepts.
However, it is important to highlight that the execution times are not re-
ally relevant for this evaluation. They depend mainly on the complexity of the
SPARQL queries required for the data extraction as well as on the availability
of SPARQL endpoints or Google App Engine servers. The goal of our evaluation
was to prove that the LDVM can be applied to different datasets providing dif-
ferent data visualizations. All these visualization examples are available in the
website and it easy to create new ones.
6 Conclusions and Future Work
We have presented the Linked Data Visualization Model (LDVM) that can be ap-
plied to create visualizations of RDF data. It allows to connect different datasets,
different data extractions and different visualizations in a dynamic way. Applying
this model, developers and designers can get a better understanding of the vi-
sualization process with data stages, transformations and operators. This model
can offer user guidance on how to create visualizations for RDF data.
We have implemented the model into an application where it is possible to
create generic visualizations for RDF. In our implementation we provide visual-
izations that support the overview task proposed by Shneiderman in his visual
seeking mantra. These visualizations are useful for getting a broad view of the
datasets, their main types, properties and the relationships between them.
Our future work focuses on complementing the Linked Data Visualization
Model with an ontology. A visualization ontology can help during the matching
process between data and visualizations, capture the intermediate data struc-
tures that can be generated and choose the visualizations more suitable for each
data structure.
Regarding our implementation of the model, we plan to develop new data
extractions and visualizations for them, e.g. charts or timelines. Having more
data extractions and visualizations will prove that the LDVM can be widely
applied.
Finally, based on the LDVM we want to develop a dashboard for visualiz-
ing and interacting with linked data through different visualizations and other
components. However, other Information Architecture components [27] such as
facets or breadcrumbs are necessary to support the other tasks for data analysis
and exploration.
References
1. Aba-Sah Dadzie and Matthew Rowe. Approaches to visualising linked data: A
survey. Semantic Web, 2(2):89–124, 2011.
2. Vladimir Geroimenko and Chaomei Chen, editors. Visualizing the Semantic Web.
Springer, 2002.
3. Stuart K. Card, J. D. Mackinlay, and Ben Shneiderman. Readings in Information
Visualization: Using Vision to Think. Academic Press, London, 1999.
4. Jeffrey Heer, Michael Bostock, and Vadim Ogievetsky. A tour through the visual-
ization zoo. Queue, 8(5):20:20–20:30.
5. Ed H. Chi. A taxonomy of visualization techniques using the data state reference
model. In Proceedings of the IEEE Symposium on Information Vizualization 2000,
INFOVIS ’00, pages 69–, Washington, DC, USA, 2000. IEEE Computer Society.
6. Uldis Bojars, Alexandre Passant, Frederick Giasson, and John Breslin. An ar-
chitecture to discover and query decentralized RDF data. volume 248 of CEUR
Workshop Proceedings ISSN 1613-0073, June 2007.
7. Tim Berners-lee, Yuhsin Chen, Lydia Chilton, Dan Connolly, Ruth Dhanaraj,
James Hollenbach, Adam Lerer, and David Sheets. Tabulator: Exploring and ana-
lyzing linked data on the semantic web. In In Proceedings of the 3rd International
Semantic Web User Interaction Workshop, 2006.
8. Experimenting with explorator: a direct manipulation generic rdf browser and
querying tool. In Workshop on Visual Interfaces to the Social and the Seman-
tic Web (VISSW2009), February 2009.
9. J.M. Brunetti, R. Gil, and R. Garcia. Facets and pivoting for flexible and usable
linked data exploration. Crete, Greece, May 2012.
10. Tuukka Hastrup, Richard Cyganiak, and Uldis Bojars. Browsing linked data with
fenfire, 2008.
11. Adapting Graph Visualization, Flavius Frasincar, Ru Telea, and Geert jan Houben.
Adapting graph visualization techniques for the visualization of rdf data. In of RDF
data, Visualizing the Semantic Web, 2006, pages 154–171, 2005.
12. David Karger and MC Schraefel. The pathetic fallacy of RDF. Position Paper for
SWUI06, 2006.
13. Magnus Stuhr, Roman Dumitru, and David Norheim. LODWheel - JavaScript-
based Visualization of RDF Data -, pages 1–12. 2011.
14. Claus Stadler, Jens Lehmann, Konrad Höffner, and Sören Auer. Linkedgeodata:
A core for a web of spatial open data. Semantic Web Journal, 2011.
15. Maria Cristina Ferreira de Oliveira and Haim Levkowitz. From visual data explo-
ration to visual data mining: A survey. IEEE Transactions on Visualization and
Computer Graphics, 9:378–394, 2003.
16. S. K. Card and J. Mackinlay. The structure of the information visualization design
space. In Proceedings of the 1997 IEEE Symposium on Information Visualiza-
tion (InfoVis ’97), INFOVIS ’97, pages 92–, Washington, DC, USA, 1997. IEEE
Computer Society.
17. E. H. H. Chi, P. Barry, J. Riedl, and J. Konstan. A spreadsheet approach to infor-
mation visualization. In Proceedings of the 1997 IEEE Symposium on Information
Visualization (InfoVis ’97), INFOVIS ’97, pages 17–, Washington, DC, USA, 1997.
IEEE Computer Society.
18. Ben Shneiderman. The eyes have it: A task by data type taxonomy for information
visualizations. In IEEE Visual Languages, number UMCP-CSD CS-TR-3665, pages
336–343, College Park, Maryland 20742, U.S.A., 1996.
19. Niklas Elmqvist and Jean-Daniel Fekete. Hierarchical aggregation for informa-
tion visualization: Overview, techniques, and design guidelines. IEEE Trans. Vis.
Comput. Graph., 16(3):439–454, 2010.
20. Ben Shneiderman. Tree visualization with tree-maps: 2-d space-filling approach.
ACM Trans. Graph., 11(1):92–99, January 1992.
21. Taowei David Wang and Bijan Parsia. Cropcircles: Topology sensitive visualization
of owl class hierarchies. In In Proc. of the 5th International Conference on Semantic
Web, pages 695–708, 2006.
22. Catherine Plaisant, Jesse Grosjean, and Benjamin B. Bederson. Spacetree: Sup-
porting exploration in large node link tree, design evolution and empirical evalua-
tion. In Proceedings of the IEEE Symposium on Information Visualization (Info-
Vis’02), INFOVIS ’02, pages 57–, Washington, DC, USA, 2002. IEEE Computer
Society.
23. Natalia Andrienko. Interactive visual clustering of large collections of trajectories.
In Working Notes of the LWA 2011 - Learning, Knowledge, Adaptation, 2011.
24. Heidrun Schumann and Christian Tominski. Analytical, visual and interactive
concepts for geo-visual analytics. J. Vis. Lang. Comput., 22(4):257–267, August
2011.
25. Sören Auer, Christian Bizer, Georgi Kobilarov, Jens Lehmann, and Zachary Ives.
Dbpedia: A nucleus for a web of open data. In In 6th Int?l Semantic Web Confer-
ence, Busan, Korea, pages 11–15. Springer, 2007.
26. Michael Bostock, Vadim Ogievetsky, and Jeffrey Heer. D3: Data-driven documents.
IEEE Trans. Visualization & Comp. Graphics (Proc. InfoVis), 2011.
27. Josep Maria Brunetti and Roberto García. Information architecture automatiza-
tion for the semantic web. In Proceedings of the 13th IFIP TC 13 international
conference on Human-computer interaction - Volume Part IV, INTERACT’11,
pages 410–413, Berlin, Heidelberg, 2011. Springer-Verlag.
... Applying information visualization techniques to the Semantic Web helps users to explore large quantities of data and interact with them. The main objectives of information visualization are to transform data into and present data as a visual representation in such a way that users can obtain a better understanding of it 15 . ...
... High level LDVM overview15 DSRM describes the visualization process in a generic way15 . ...
... High level LDVM overview15 DSRM describes the visualization process in a generic way15 . ...
Chapter
Full-text available
The main subject of the work is the visualization of typed links in Linked Data. The academic subjects relevant to the paper in general are the Semantic Web, the Web of Data and information visualization. The Semantic Web, invented by Tim Berners-Lee in 2001, was announced as an extension to the World Wide Web (Web 2.0). The actual area of investigation concerns the connectivity of information on the World Wide Web. To be able to explore such interconnections, visualizations are critical requirements as well as a major part of processing data in themselves. In the context of the Semantic Web, representation of information interrelations can be achieved using graphs. The aim of the article is to primarily describe the arrangement of Linked Data visualization concepts by establishing their principles in a theoretical approach. Putting design restrictions into context leads to practical guidelines. By describing the creation of two alternative visualizations of a commonly used web application representing Linked Data as network visualization, their compatibility was tested. The application-oriented part treats the design phase, its results, and future requirements of the project that can be derived from this test. VISUALISIERUNG VON TYPISIERTEN LINKS IN LINKED DATA Zusammenfassung: Das Themengebiet der Arbeit behandelt Visualisierungen von typisierten Links in Linked Data. Die wissenschaftlichen Gebiete, die im Allgemeinen den Inhalt des Beitrags abgrenzen, sind das Semantic Web, das Web of Data und Informationsvisualisierung. Das Semantic Web, das von Tim Berners Lee 2001 er-funden wurde, stellt eine Erweiterung zum World Wide Web (Web 2.0) dar. Aktuelle Forschungen beziehen sich auf die Verknüpfbarkeit von Informationen im World Wide Web. Um es zu ermöglichen, solche Verbindungen wahrnehmen und verarbeiten zu können sind Visualisierungen die wichtigsten Anforderungen als Hauptteil der Daten-verarbeitung. Im Zusammenhang mit dem Sematic Web werden Repräsentationen von zuhammenhängenden Informationen anhand von Graphen gehandhabt. Der Grund des Entstehens dieser Arbeit ist in erster Linie die Beschreibung der Gestaltung von Linked Data-Visualisierungskonzepten, deren Prinzipien im Rahmen einer theo-retischen Annäherung eingeführt werden. Anhand des Kontexts führt eine schrittweise Erweiterung der Informationen mit dem Ziel, praktische Richtlinien anzubieten, zur Mitteilungen der VÖB 70 (2017) Nr. 2: Metadata-Metadaten 180
... Applying information visualization techniques to the Semantic Web helps users to explore large quantities of data and interact with them. The main objectives of information visualization are to transform data into and present data as a visual representation in such a way that users can obtain a better understanding of it 15 . ...
... High level LDVM overview15 DSRM describes the visualization process in a generic way15 . ...
... High level LDVM overview15 DSRM describes the visualization process in a generic way15 . ...
Article
Full-text available
Das Themengebiet der Arbeit behandelt Visualisierungen von typisierten Links in Linked Data. Die wissenschaftlichen Gebiete, die im Allgemeinen den Inhalt des Beitrags abgrenzen, sind das Semantic Web, das Web of Data und Informationsvisualisierung. Das Semantic Web, das von Tim Berners Lee 2001 erfunden wurde, stellt eine Erweiterung zum World Wide Web (Web 2.0) dar. Aktuelle Forschungen beziehen sich auf die Verknüpfbarkeit von Informationen im World Wide Web. Um es zu ermöglichen, solche Verbindungen wahrnehmen und verarbeiten zu können sind Visualisierungen die wichtigsten Anforderungen als Hauptteil der Datenverarbeitung. Im Zusammenhang mit dem Sematic Web werden Repräsentationen von zuhammenhängenden Informationen anhand von Graphen gehandhabt. Der Grund des Entstehens dieser Arbeit ist in erster Linie die Beschreibung der Gestaltung von Linked Data-Visualisierungskonzepten, deren Prinzipien im Rahmen einer theoretischen Annäherung eingeführt werden. Anhand des Kontexts führt eine schrittweise Erweiterung der Informationen mit dem Ziel, praktische Richtlinien anzubieten, zur Vernetzung dieser ausgearbeiteten Gestaltungsrichtlinien. Indem die Entwürfe zweier alternativer Visualisierungen einer standardisierten Webapplikation beschrieben werden, die Linked Data als Netzwerk visualisiert, konnte ein Test durchgeführt werden, der deren Kompatibilität zum Inhalt hatte. Der praktische Teil behandelt daher die Designphase, die Resultate, und zukünftige Anforderungen des Projektes, die durch die Testung ausgearbeitet wurden.
... e client layer enables users to interact with the system using retrieval and visualization tools. e coherent LOD visualization is an essential part of any LODbased system to enable users to utilize the Web of Data and enhance the accessibility and usability of a system [36][37][38]. ...
Article
Full-text available
The COVID-19 data is critical to support countries and healthcare organizations for effective planning and evidence-based practices to counter the pressures of cost reduction, improved coordination, and outcome and produce more with less. Several COVID-19 datasets are published on the web to support stakeholders in gaining valuable insights for better planning and decision-making in healthcare. However, the datasets are produced in heterogeneous proprietary formats, which create data silos and make data discovery and reuse difficult. Further, the data integration for analysis is difficult and is usually performed by the domain experts manually, which is time-consuming and error-prone. Therefore, an explicit, flexible, and widely acceptable methodology to represent, store, query, and visualize COVID-19 data is needed. In this paper, we have presented the design and development of the Linked Open COVID-19 Data system for structuring and transforming COVID-19 data into semantic format using explicitly developed ontology and publishing on the web using Linked Open Data (LOD) principles. The key motivation of this research is the evaluation of LOD technology as a potential option and application of the available Semantic Web tools (i.e., Protégé, Excel2RDF, Fuseki, Silk, and Sgvizler) for building LOD-based COVID-19 information systems. We have also underpinned several use-case scenarios exploiting the LOD format of the COVID-19 data, which could be used by applications and services for providing relevant information to the end-users. The effectiveness of the proposed methodology and system is evaluated using the system usability scale and descriptive statistical methods and results are found promising.
... However, most of the work (e.g. [3,5,18,31]) focuses on Linked Data visualisation to experts and not specifically on informed consent visualisation for non-expert users. This topic is the focus of [9] and [10]. ...
Conference Paper
The GDPR legislation has brought to light one’s rights and has highlighted the importance of consent, which has caused a major shift in how data processing and sharing are handled. Data sharing has been a popular research topic for many years, however, a unified solution for the transparent implementation of consent, in compliance with GDPR that could be used as a standard, has not been presented yet. This research proposes a solution for implementing informed consent for sensor data sharing in compliance with GDPR with semantic technology, namely knowledge graphs. The main objectives are to model the life cycle of informed consent (i.e. the request, comprehension, decision and use of consent) with knowledge graphs so that it is easily interpretable by machines, and to graphically visualise it to individuals in order to raise legal awareness of what it means to consent and the implications that follow.
... which enables retrieval of large amounts of data in a uniform way without the need to know the database schemas; and Maelstrom which offers guidance to document and disseminate study metadata across collaborating institutions [4]. Regarding Linked Data (LD), various works have been proposed in the literature [5][6][7]. Besides, several tools offering RDF and linked data visualization have been developed, e.g., Sgvizler [8], LODWheel [9], IsaViz [10]. RDF-Gravity [11], RML [12], etc. ...
Article
Full-text available
Given the huge amount of heterogeneous data stored in different locations, it needs to be federated and semantically interconnected for further use. This paper introduces WINFRA, a comprehensive open-access platform for semantic web data and advanced analytics based on natural language processing (NLP) and data mining techniques (e.g., association rules, clustering, classification based on associations). The system is designed to facilitate federated data analysis, knowledge discovery, information retrieval, and new techniques to deal with semantic web and knowledge graph representation. The processing step integrates data from multiple sources virtually by creating virtual databases. Afterwards, the developed RDF Generator is built to generate RDF files for different data sources, together with SPARQL queries, to support semantic data search and knowledge graph representation. Furthermore, some application cases are provided to demonstrate how it facilitates advanced data analytics over semantic data and showcase our proposed approach toward semantic association rules.
... In addition, one can interact with the representation so that one can interpret it more effectively. One created a pipeline ( Figure 2) and used the Date State Reference Model (DSRM), proposed by Chi [59], together with the Linked Data Visualization Model (LDVM), proposed by Brunetti, Auer and García [60], as inspiration. ...
... and Maelstrom (www.maelstrom-research.org). In the context of RDF and Linked Open Data, various works have been proposed in the literatures [2,3]. In addition, several tools offering RDF and linked data visualization have been developed e.g., Sgvizler [6], LODWheel [7], IsaViz 1 . ...
Conference Paper
Full-text available
Given the increasing number of heterogeneous data stored in relational databases, file systems or cloud environment, it needs to be easily accessed and semantically connected for further data analytic. The potential of data federation is largely untapped, this paper presents an interactive data federation system (https://vimeo.com/319473546) by applying large-scale techniques including heterogeneous data federation, natural language processing, association rules and semantic web to perform data retrieval and analytics on social network data. The system first creates a Virtual Database (VDB) to virtually integrate data from multiple data sources. Next, a RDF generator is built to unify data, together with SPARQL queries, to support semantic data search over the processed text data by natural language processing (NLP). Association rule analysis is used to discover the patterns and recognize the most important co-occurrences of variables from multiple data sources. The system demonstrates how it facilitates interactive data analytic towards different application scenarios (e.g., sentiment analysis, privacy-concern analysis, community detection).
Chapter
Increasing numbers of Linked Open Datasets are being published, and many possible data visualisations may be appropriate for a user’s given exploration or analysis task over a dataset. Users may therefore find it difficult to identify visualisations that meet their data exploration or analyses needs. We propose an approach that creates conceptual models of groups of commonly used data visualisations, which can be used to analyse the data and users’ queries so as to automatically generate recommendations of possible visualisations. To our knowledge, this is the first work to propose a conceptual modelling approach to recommending visualisations for Linked Data.
Article
Recent years have witnessed progress of public institutions in making their datasets available online, free of charge, for re-use. There have been however limited studies which assess the actual effectiveness of different communication media in making key facts visible to citizens. This article analysed and systematically compared two representations which are relevant in the context of open government data: geovisualizations and data tables. An empirical user study (N=16) revealed that both types of representations have their strengths: geovisualizations make spatial knowledge and the attractiveness of open government data more visible, while data tables are more adequate for the communication of numerical data. The ideas presented are relevant to open data publishers interested in strategies to effectively put the hidden knowledge in current open government datasets into the hands of citizens.
Conference Paper
Full-text available
Visualizing Resource Description Framework (RDF) data to support decision-making processes is an important and challenging aspect of consuming Linked Data. With the recent development of JavaScript libraries for data visualization, new opportunities for Web-based visualization of Linked Data arise. This paper presents an extensive evaluation of JavaScript-based libraries for visualizing RDF data. A set of criteria has been devised for the evaluation and 15 major JavaScript libraries have been analyzed against the criteria. The two JavaScript libraries with the highest score in the evaluation acted as the basis for developing LODWheel (Linked Open Data Wheel)-a prototype for visualizing Linked Open Data in graphs and charts-introduced in this paper. This way of visualizing RDF data leads to a great deal of challenges related to data-categorization and connecting data resources together in new ways, which are discussed in this paper.
Article
Full-text available
A web of linked RDF data may be enabled by standards specifying how links should be made in RDF and under what conditions they should be followed as well as powerful generic RDF browsers that can traverse an open web of RDF resources. The Tabulator is an RDF browser, which is designed both for new users to provoke interest in the Semantic Web and give them a means to access and interact with the entire web of RDF data, and for developers of RDF content to provide incentive for them to post their data in RDF, to refine and promote RDF linking standards, and to let providers see how their data interacts with the rest of the Semantic Web. A challenge for Semantic Web browsers is to bring the power of domain-specific applications to a generic program when new unexpected domains can be encountered in real time. The Tabulator project is an attempt to demonstrate and utilize the power of linked RDF data with a user-friendly Semantic Web browser that is able to recognize and follow RDF links to other RDF resources based on the user's exploration and analysis.
Article
Full-text available
The Semantic Web eases data and information integration tasks by providing an infrastructure based on RDF and ontologies. In this paper, we contribute to the development of a spatial Data Web by elaborating on how the collaboratively collected OpenStreetMap data can be interactively transformed and represented adhering to the RDF data model. This transformation will simplify information integration and aggregation tasks that require comprehensive background knowledge related to spatial features such as ways, structures, and landscapes. We describe how this data is interlinked with other spatial data sets, how it can be made accessible for machines according to the Linked Data paradigm and for humans by means of several applications, including a faceted geo-browser. The spatial data, vocabularies, interlinks and some of the applications are openly available in the LinkedGeoData project.
Conference Paper
Full-text available
The success of Open Data initiatives has increased the amount of data available on the Web. Unfortunately, most of this data is only available in raw tabular form, what makes analysis and reuse quite difficult for non-experts. Linked Data principles allow for a more sophisticated approach by making explicit both the structure and semantics of the data. However, from the end-user viewpoint, they continue to be monolithic files completely opaque or difficult to explore by making tedious semantic queries. Our objective is to facilitate the user to grasp what kind of entities are in the dataset, how they are interrelated, which are their main properties and values, etc. Rhizomer is a tool for data publishing whose interface provides a set of components borrowed from Information Architecture (IA) that facilitate awareness of the dataset at hand. It automatically generates navigation menus and facets based on the kinds of things in the dataset and how they are described through metadata properties and values. Moreover, motivated by recent tests with end-users, it also provides the possibility to pivot among the faceted views created for each class of resources in the dataset.
Article
A wealth of information has recently become available as browsable RDF data on the Web, but the selection of client applications to interact with this Linked Data remains lim-ited. We show how to browse Linked Data with Fenfire, a Free and Open Source Software RDF browser and editor that employs a graph view and focuses on an engaging and interactive browsing experience. This sets Fenfire apart from previous table-and outline-based Linked Data browsers.
Conference Paper
OWL ontologies present many interesting visualization challenges. Here we present CropCircles, a technique designed to view the class hierarchies in ontologies as trees. We place special emphasis on topology understanding when designing the tool. We drew inspiration from treemaps, but made substan- tial changes in the representation and layout. Most notably, the spacefillingness of treemap is relaxed in exchange for visual clarity. We outline the problem scape of visualizing ontology hierarchies, note the requirements that go into the de- sign of the tool, and discuss the interface and implementation. Finally, through a controlled experiment involving tasks common to understanding ontologies, we show the benefits of our design.