Schematization of Clutter Reduction Techniques in Geographic

Node-link Diagrams using Task-based Criteria

Alberto Debiasi, Bruno Sim

oes and Raffaele De Amicis

Fondazione Graphitech, Trento, Italy

Keywords:

Visual Clutter, Geographic Node-link Diagram, Edge Congestion, Geo-referenced Networks.

Abstract:

Visual clutter is a hot topic in the domain of node-link diagrams as it negatively affects usability, aesthetics

and data interpretation. The organization of items, i.e. the way nodes and links are positioned in the display,

is one problem among many that leads to visual clutter. In previous work, different techniques were proposed

to reduce the clutter that depends on the organization of nodes and links. However, a schematization of such

techniques by task was never considered. Approaching the problem by task would be more efﬁcient since

visual clutter, by deﬁnition, depends on the task to be performed. In this paper, we propose a solution to

visual clutter driven by the type of task. In particular, the aim of our work is to provide an answer to the

following question: Given a task and a geographic node-link diagram, which are the appropriated techniques

to reduce the visual clutter that depends on the spatial organization of nodes and links. In our solution, we have

classiﬁed tasks into a limited number of task groups. For each tasks group, we have identiﬁed and analyzed

issues leading to a performance degradation. The ﬁnal outcome consists on a list of good candidate techniques

for each task group. The selected techniques are the results of a survey that selects only approaches that act

on the position of nodes and links.

1 INTRODUCTION

The node-link diagram is a powerful tool for the visu-

alization of relationships between entities. However,

such visualization often suffer from visual clutter (Liu

et al., 2014; Sun et al., 2013; Ellis and Dix, 2007) af-

fecting usability, aesthetics and data interpretation.

In previous studies visual clutter is deﬁned as:

“the state in which excess items, or their represen-

tation or organization, lead to a degradation of per-

formance at some task” (Rosenholtz et al., 2005).

Hence, visual clutter depends on the task, as shown

in Figure 1. If the task is “ﬁnd a node”, only the

diagram in Figure 1(a) should be considered clut-

tered (Holten and Van Wijk, 2009; Ersoy et al., 2011;

Hurter et al., 2012) because many nodes are occluded.

If the task is “given a node, ﬁnd the connected nodes”,

also the diagram in Figure 1(b) should be considered

cluttered (Wong et al., 2003; Wong and Carpendale,

2007; Schmidt et al., 2010) because ambiguities are

present. If the task is considered of high-level, such

as “understand the story described by the data”, then

the diagram in Figure 1(c) should also be considered

cluttered (Phan et al., 2005; Verbeek et al., 2011; De-

biasi et al., 2014) as it is not aesthetically pleasing.

In this work we focus on the problem of organiza-

tion of items, i.e. the way nodes and links are posi-

tioned in the display, for which we propose a solution

that is driven by task group. In geographic node-link

diagrams, this aspect is critical because the position

of nodes is ﬁxed accordingly to geographical infor-

mation. As opposite, we do not examine clutter that

depends on the graphical representation of items, for

example when the color of the nodes is the same of the

background map, or when size of nodes is too small

or too large taking into account the display size and

the number of items.

Our work answers to the following research ques-

tion: Given a task and a geographic node-link dia-

gram, which are the appropriated techniques to reduce

the visual clutter that depends on the spatial organiza-

tion of nodes and links.

We start identifying and analyzing the main prob-

lems that lead to a degradation of tasks performance,

i.e. uninterpretable representation, occlusion, ambi-

guity, and unaesthetic representation. Then, we di-

vide the tasks into task groups, and for each group we

associate the problems that characterize it. We use

the task taxonomy for graph visualization (Lee et al.,

2006) that covers exploratory and analytical tasks. We

include a tasks group related to the aesthetic of the

visualization, i.e. explanatory tasks. Finally, we pro-

Debiasi, A., Simões, B. and Amicis, R.

Schematization of Clutter Reduction Techniques in Geographic Node-link Diagrams using Task-based Criteria.

DOI: 10.5220/0005674801070114

In Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2016) - Volume 2: IVAPP, pages 109-116

ISBN: 978-989-758-175-5

109

(a)

(b) (c)

Figure 1: Examples of different visual clutter in geographic

node-link diagrams.

vide a list of good candidate techniques for each prob-

lem.

The selected techniques are the results of a survey

that selects only approaches that act on the position of

nodes and links. We restrict the analysis to this sub-

set mainly because the position is the most used vari-

able in the context of geographical node-link diagram

to reduce clutter. We do not consider force-directed

techniques on nodes because they do not preserve spa-

tial information.

This paper is structured as follows. First we sum-

marize, in Section 2, previous works on visual clut-

ter in geographic node-link diagrams. In Section 3,

we identify the visual clutter problems. Then we de-

ﬁne in Section 4 our classiﬁcation of tasks by groups.

Lastly, in Section 5 we itemize the main clutter reduc-

tion techniques according to the deﬁned criteria. An

overview of the results obtained and of future work is

presented in Section 6.

2 EXISTING SURVEYS

In literature different surveys on clutter reduction

techniques are presented. However, they do not con-

sider the different problems related to visual clutter

(e.g. occlusion and ambiguity), nor they take into ac-

count the task.

In (Sun et al., 2013; Liu et al., 2014) visual analyt-

ics techniques were presented and a section was ded-

icated to the clutter reduction methods in large graph

layouts. However, no hints are made with respect to

the aforementioned aspects. The clutter problem and

the related solutions are also mentioned in different

categorizations of node-link diagrams (Hadlak et al.,

2015; Debiasi et al., 2015a), but not as main chal-

lenge.

Zhou et al. (Zhou et al., 2013) provide a survey

on edge bundling techniques. They described the al-

gorithms accordingly with the way they generate the

ﬁnal graph layout, i.e. they distinguished cost-based,

geometry based and image-based techniques. As op-

posite, we classify the different approaches accord-

ingly with the criteria they fulﬁll, by looking at the

ﬁnal layout they generate. In the survey (Tominski

et al., 2014), lens techniques in the context of visu-

alization are categorized according to data types and

tasks. Although many clutter reduction techniques are

included, they are classiﬁed according with the pro-

prieties of the solutions.

A previous survey (Ellis and Dix, 2007) focused

on cluttered visualizations caused by huge amounts

of data. The authors classiﬁed clutter reduction meth-

ods deﬁning eight criteria based on their experience

and on the study of the related literature. The main

difference between our work and theirs is that their

criteria focus on the techniques. We provide crite-

ria derived from an analysis of cluttered layouts and

tasks. Moreover, our work focus speciﬁcally on inap-

propriate organization of nodes and links in node-link

diagrams and not on clutter caused by huge data.

3 VISUAL CLUTTER PROBLEMS

Focusing on the organization of items, clutter is

mainly the result of the overlapping (overplotting or

overdrawing) of those items, i.e. items rendered on

top (on near) of each other. In this work, the consid-

ered items are nodes, links (straight lines or curves),

and the geographical surface (background map). We

deﬁne the symbology (element1, element2) to indi-

cate that “element1 is rendered on top of element2”.

We describe four problems associated to clutter: unin-

terpretable representation, occlusion, ambiguity, and

unaesthetic representation.

3.1 Problem of Uninterpretability

The readability of node-link diagrams deteriorates

when items are badly located or when the size of the

graph and its link density increase (Ghoniem et al.,

2004). The problem of uninterpretable representa-

tion occurs when items become impossible to iden-

tify, making the visualization useless.

IVAPP 2016 - International Conference on Information Visualization Theory and Applications

110

Table 1: Occlusion of elements (in red) results in loss of

information. For each scenario, we apply an approach that

acts on elements opacity to solve the problem.

Element on back

Node Link Geo-Surface

Element on top

Node Link

3.2 Problem of Occlusion

Table 1 shows information loss of geographic node-

link diagram taking into account all possible combi-

nations between elements.

Here some elements are hidden behind other ele-

ments:

• (node,node);(node,link);(node,geo-surface):

Nodes located is the same spatial location or near

with each other may cause their occlusion. In

the same way, nodes can occlude also links or

background map.

• (link,node);(link,link);(link,geo-surface): Links

obscure nodes, links, information in the back-

ground, map labels or map features. This prob-

lem is accentuated with thick links, or with a large

number of links.

3.3 Problem of Ambiguity

Table 2 shows the ambiguity problem of node-link di-

agram taking into account different cases of ambigu-

ity:

• (node,link);(link,node): It can be difﬁcult to as-

sess, when a link crosses a node, or vice-versa,

whether this link is incident to this node or it is

merely crossing it. This makes it difﬁcult to iden-

tify the actual sources and destinations of links. If

nodes are rendered on bottom of links, links may

obscure the content of the nodes. In the opposite

case, the unambiguous layout candidates are more

in numbers.

• (link,link): Link crossings are an important fac-

tor in readability (Purchase et al., 2004). We con-

sider colliding links also when they are very close

to each other and are almost parallel. When eyes

try to follow a link to its destination, small cross-

ing angles between this link and other links cre-

ate multiple paths along the direction of the eye

Table 2: Ambiguity cases and their possible unambiguous

conﬁgurations.

Ambiguous Layout

Unambiguous Configurations

(

node,link): node crosses link

(

link,node): link crosses node

(

link,link): link crosses link

(

𝑙𝑖𝑛𝑘, 𝑙𝑖𝑛𝑘)

: link overlaps link

(

𝑙𝑖𝑛𝑘, 𝑙𝑖𝑛𝑘)

: link overlaps link - limit case

….

out

plane

train

….

movement, either taking eyes to the wrong path,

or slowing down the eye movement. However, if

two links share one node, this ambiguity does not

occur.

• (link,link)

: In this case links overlap each other.

Hence, it can be considered an occlusion crite-

ria. This happen if they have one or both incident

nodes in common. For example if links repre-

sent path segments, two or more links completely

overlaps if they have same origin and destination

but different time period.

3.4 Problem of Unaesthetic

The problem of unaesthetic representation differs

from the occlusion problem, because we do not neces-

sarily have loss of information. As shown in Table 3,

partial overlapping causes a decrement in visual qual-

ity of the layout. Thus, for the aesthetic of the repre-

sentation, the following cases are relevant to describe

the effect of clutter:

• (node,node): Nodes located is the same spatial lo-

cation or near with each other may cause partial

overlapping of nodes.

• (link,node);(node,link): Links can cut directly

across a node (or vice-versa) interfering with the

visual quality of the layout.

• (link,link): It is possible to have links that are

close to or actually overlapping each other.

• (link,geo-surface);(node,geo-surface): When

links or nodes overlap a region, they may

decrease the layout aesthetic.

Schematization of Clutter Reduction Techniques in Geographic Node-link Diagrams using Task-based Criteria

111

Table 3: Partial overlapping of elements (colored in red)

causes a decrement of layout aesthetic. For each case an

approach acting on elements position is applied to solve the

problem.

Element on back

Node Link Geo-Surface

Element on top

Node Link

3.5 Criteria Fulﬁllment

We use the criteria mentioned below, to valuate the

candidate solutions for the identiﬁed problems.

The problems of uninterpretable representation

and occlusion, are analyzed by applying each tech-

nique to example in Figure 1(a). Although such dia-

gram does not present a background image, we enrich

the visualization with the appropriate map. For ambi-

guity problems in Table 2, each technique is applied

to example in Figure 1(b). For the problem in Table 3,

each technique is applied to example in Figure 1(c).

Although such diagram does not present overlapping

nodes, we enrich the visualization increasing the size

of the nodes.

For each problem, a technique is marked as:

• 3: if the technique removes the problem.

• 3

∗

: if the technique may removes the problem,

however, this condition is not guaranteed for all

the cases.

• -: if the technique does not provide any evidence

that it reduces or removes the problem.

• 7: if the technique increase the problem. In such

cases, the goal of the technique is another.

4 TASKS GROUPS FOR

GEOGRAPHIC NODE-LINK

DIAGRAMS

We distinguish different tasks groups in the context

of geo-referenced networks, to better understand vi-

sual clutter. At high level of abstraction, we ap-

ply the general division for information visualization

tasks (Keim et al., 2006):

• Exploratory and Analytical tasks: Visual explo-

ration allows the possibility to get new insight. In

case of analytical scenario, the user knows always

the task, being it implicit or explicit. We classify

the tasks identiﬁed in (Lee et al., 2006) as follows:

– Graph-speciﬁc tasks: The tasks are related to

the topology of the graph, i.e. “given a node,

ﬁnd adjacent nodes” and “given a link, ﬁnd in-

cident nodes”.

– General low-level tasks: The user examines

each item of the network to make new discover-

ies. The tasks are: “ﬁnd item”, “retrieve value”,

“ﬁlter item”, etc.

– Overview tasks: Some high-level tasks require

only an overview of the graph such as ﬁnding

clusters of related nodes, ﬁnding patterns and

outliers.

• Explanatory tasks: The main goal is to make sense

(i.e. associative thinking) of a story visually de-

scribed by the data. In literature, common ex-

amples of node-link diagram for such tasks are

ﬂow maps: geographical maps where straight or

curved lines represent the movement of groups of

objects from one location to another. The thick-

ness of the line identiﬁes the number of moving

objects, see Figure 1(c).

4.1 Analysis of Tasks Groups and

Clutter Problems

In this section we identify the problems that charac-

terize each tasks group, as shown in Figure 2.

In the task group “Overview Tasks”, visual clutter

is interconnected to the interpretation of the represen-

tation. In these tasks it is not essential to solve the

problem of occlusion, ambiguity and unaesthetic.

In the task group “General low-level tasks”, visual

clutter involve also the occlusion of items. It becomes

difﬁcult to understand information encoded in visual

variables of occluded elements.

In the task group “Graph speciﬁc tasks” ambiguity

is an important aspect to take care of. In this task

group, we are not only searching for hidden elements,

but also we are trying to have a clearer (unambiguous)

layout.

In the task group “Associative Visualization

Tasks” there is the problem of unaesthetic representa-

tion. Due to the layout, some partial overlap between

elements may occurs, which cause a decrement in vi-

sual quality of the layout. The aesthetic is not a prior-

ity in exploratory and analytical tasks.

IVAPP 2016 - International Conference on Information Visualization Theory and Applications

112

Graph-specific

Tasks

TASKS

General low-level

Tasks

Overview

Tasks

Ambiguity

Occlusion

VISUAL CLUTTER PROBLEMS ON THE USER

Uninterpretable

representation

Unaesthetic

representation

Associative

Visualization Tasks

Explanatory Tasks Exploratory and Analytical Tasks

Figure 2: Task groups are affected by different types of vi-

sual clutter.

5 APPLIED SCHEMATIZATION

Each technique surveyed in this work is compared

with respect to our criteria, see Table 4. For each task

group, analyzing the table, we propose the following

candidate solutions.

5.1 Overview Tasks

Edge bundling is one of the main techniques that

makes a layout more easy to interpret. In edge

bundling, links of a graph are bundled together ac-

cording to deﬁned conditions. In the ﬁrst generation

of such algorithms (Qu et al., 2007; Zhou et al., 2008;

Telea and Ersoy, 2010; Holten, 2006; Cui et al., 2008;

Holten and Van Wijk, 2009; Gansner et al., 2011;

Lambert et al., 2010a), without considering the com-

bination with other techniques, the main beneﬁt is the

reduction of the number of visible items.

5.2 General Low-level Tasks

Winding Roads (Lambert et al., 2010b),

KDEEB (Hurter et al., 2012), and

ADEB (Peysakhovich et al., 2015), when com-

pared to previous Edge Bundling approaches, are

able to avoid overlapping of links with nodes and map

portions. Edge Bundling techniques are also designed

to manage multivariate networks. DEB (Selassie

et al., 2011) is able to separate opposite-direction

bundles, emphasizing the graph structure. SBEB (Er-

soy et al., 2011) bundled similar links according to

directions and also further variables associated to

links.

MoleView (Hurter et al., 2011) is a semantic lens

that selects a set of data elements located within the

lens’ radius and having an attribute value deﬁned by

the user. In case of graph layout the lens moves the

control points that compose links around the lens.

EdgeAnalyzer (Panagiotidis et al., 2011) is able to

select speciﬁc sub-groups of links on dense areas.

Those techniques are able to reveal nodes, links and

map occluded by the links.

5.3 Graph-speciﬁc Tasks

Edge bundling techniques can be designed to reduce

ambiguities. SideKnot (Peng et al., 2012) focused

on (link,link) ambiguity. Although it does not re-

duce the ambiguity, with respect to the aforemen-

tioned bundling techniques, it does not even increase

that problem. Stub Bundling (Nocaj and Brandes,

2013) made a step forward. It uses parallel routing

of links to facilitate the display of additional data at-

tributes by varying width or color. Hence, it solves

the (link,link)

ambiguity problem. Ambiguity-Free

Edge-Bundling (Luo et al., 2012) is an approach that

bundles only links which share a common node. In

this way the (link,link) ambiguity in not increased dur-

ing the bundling procedure. Moreover, this method

reroutes links that pass over nodes reducing the

(node,link) ambiguity. Finally, Edge Routing with Or-

dered Bundles (Pupyrev et al., 2012) is able to fulﬁll

most of our criteria. Here, links are placed in parallel

channels to avoid overlaps. This approach could the-

oretically be used to create ﬂow maps, however, the

result has to be evaluated in terms of aesthetic.

Links can be drawn as curves in a 3D space to

reduce (link,link) and (node,link) ambiguities, and

nodes are snapped over a geographical surface to pre-

serve the spatial context (Cox et al., 1996; Munzner

et al., 1996). Curved links allow more display space

compared to their straight counterparts and poten-

tially reduce visual clutter (Xu et al., 2012).

Interactive Bundling (Riche et al., 2012) is an in-

teractive technique that generates crossing-minimal

bundles that are routed to distinguish them. Wong et

al. (Wong et al., 2003) proposed EdgeLens; a tech-

nique that iteratively curves graph links away from

the point of focus. This consents to disambiguate the

relationship between nodes and links without losing

information. An analogous multi-touch technique is

PushLens (Schmidt et al., 2010). 3DArcLens (Debi-

asi et al., 2015b) extends the functionalities of Edge-

Lens, distinguishing the distorted links around the

lens. As drawback, the distorted links may cross with

the surrounding lines causing further ambiguity. With

Edge Plucking the user can drag groups of links away

to clarify cluttered zones and specify links or nodes

to be left unmoved (Wong and Carpendale, 2007).

However, Edge Plucking requires a certain amount

of manual effort. Bearing this in mind, Schmidt et

al. (Schmidt et al., 2010) designed (but not imple-

mented) multi-touch interaction techniques based on

Schematization of Clutter Reduction Techniques in Geographic Node-link Diagrams using Task-based Criteria

113

Table 4: Techniques to reduce unaesthetic representation/ambiguity/occlusion/uninterpretable in geographic node-link dia-

gram. From top to bottom, the color identiﬁes the task group: Overview Tasks, General Low-level Tasks, Graph-speciﬁc

Tasks, and Associative Visualization Tasks.

Clutter Reduction

Techniques

Unaesthetic Representation Ambiguity Occlusion

Uninterpretable

Representation

(link, link)

(node, link)

(link, node)

(link, map)

(node, node)

(node, map)

(link, link)

(node, link)

(link, node)

(link, link)

(node, link)

(link, node)

(link, map)

(node, node)

(node, map)

Edge Bundlings (1° Generation)

- - - - -     * * - 

SBEB, DEB

- - - - -  *  - * * - 

Winding Road, KDEEB, ADEB

-  - - -       - 

SideKnot

- - - - - -   - * * - 

MoleView, EdgeAnalyzer

- * -    - -

3DArcLens

* *     - -

3D curving edges

* -  * * * - -

Interactive Bundling

 - - -   - -

Interactive Link Fanning

-  - - - - - -

Link Magnet, Interactive Link Legend

- * - - - - - -

EdgeLens, PushLens

- -     - -

MultiTouch Techniques

 -     - -

Stub Bundling

* - - - - -   - * * - -

Ambiguity-Free Edge-Bundling

-  - - - - -  -  * - -

Edge Routing with Ordered Bundles

   - -    -   - -

Flow Map Layout

* * - - - - - - - - - - -

Supervised Flow Map Layout

  - - - - - - - - - - -

Conuent Spiral Drawings,

Flow Map Layout via Spiral Trees

   - - - - - - - - - -

Necklace Maps

- * -  * - - - - - - -

link displacement.

Other interactive lenses were designed to reduce

clutter accordingly to the multivariate encoding of

links (Riche et al., 2012). For example, Interactive

Link Fanning creates space between links incident

to a selected node, to show labels or arrowheads for

individual links. With Link Magnet, when a visual

object representing a data attribute is dragged, data

items are attracted by an amount depending on the

attribute value of the item. Finally, with Interactive

Link Legends, link curvature encodes semantic infor-

mation such as different types of links in heteroge-

neous graphs.

5.4 Associative Visualization Tasks

In Necklace Map (Speckmann and Verbeek, 2010),

nodes are rearranged in circular layouts to remove the

overlapping between them. However, the spatial con-

text of nodes is preserved for two reasons: the nodes

are moved not too far from their original position and

the color of nodes is used to associate them with their

original location on the background map.

Phan, et al. (Phan et al., 2005) developed a method

to generate a ﬂow map layout in a recursive and sim-

ple manner. The overlapping between links and nodes

is only partially solved because undesired crossings

may occur. However, in a post processing phase, user

has the possibility to modify ﬂow lines by moving

their control points. Debiasi et al. (Debiasi et al.,

2014) presented a method to generate ﬂow map lay-

outs using a force directed approach to remove the

(link-node),(node-node) overlapping.

A second generation of ﬂow map algorithms was

developed to satisfy all the criteria related to the par-

tial overlapping of links. Verbeek, et al. (Verbeek

et al., 2011) introduced a method using spiral trees,

i.e. links are logarithmic spirals, implemented as cu-

bic Hermite splines. A different approach called Con-

ﬂuent Spirals (Nocaj and Brandes, 2013) consists of

smooth drawings in which link direction are repre-

sented by increasing curvature.

IVAPP 2016 - International Conference on Information Visualization Theory and Applications

114

6 CONCLUSIONS

This work provides a list of criteria to classify the

effects of the visual clutter on geographic node-link

diagrams on different scenarios. The scope of this

work is not the creation of a rank of techniques, but

a classiﬁcation that helps the reader to decide, given

a task, on the list of candidate solutions that help to

reduce the clutter in a geographical node-link dia-

gram. Moreover, it provides guidelines to the design

of novel techniques, helping the researchers to focus

on a well-deﬁned list of criteria to fulﬁll. As shown in

Table 4, among the techniques we surveyed there are

no solution that are capable of satisfying all criteria.

Regarding edge bundling techniques, the fulﬁll-

ment of criteria depends on the information used.

Starting from unambiguous layout, aggregating links

with no information about nodes increase all the

ambiguity cases. Information about their incident

nodes is needed to solve the ambiguity of overlapping

links. As opposite, information about crossing links

is needed to avoid completely the link-link ambiguity.

Finally, information about nodes positions is needed

to avoid crossing among links and nodes.

It is possible that some techniques that remove or

reduce the occlusion of items affect negatively the

graph-based tasks. The combined use of techniques

that act on position of nodes and links with tech-

niques that act on other visual variables can further

improve the ﬁnal result. From the proposed classiﬁ-

cation only one approach (Necklace Map) satisﬁes the

partial overlapping criteria related to occluding nodes.

The reason is the difﬁculties in rearranging the nodes

without losing their geographic information.

As future work we plan to extend this schema-

tization into a classiﬁcation of techniques, based on

their intent, e.g. “put in parallel”, “aggregate”, “push

away”. Furthermore, we could include techniques

that act on other visual variables such as color, ﬁnal

image rendering, etc. Finally, the task classiﬁcation

has to be improved. No group takes into account the

background map in their tasks.

ACKNOWLEDGEMENTS

This research has been supported by the Euro-

pean Commission under the c-Space (G.A. 611040)

and the LIFE+IMAGINE (LIFE12/ENV/IT/001054)

projects. It has been carried on in the context of the

National Geoportal project for the Italian Ministry of

Environment.The authors are solely responsible this

work.

REFERENCES

Cox, K. C., Eick, S. G., and He, T. (1996). 3d geographic

network displays. ACM Sigmod Record, 25(4):50–54.

Cui, W., Zhou, H., Qu, H., Wong, P. C., and Li, X. (2008).

Geometry-based edge clustering for graph visualiza-

tion. Visualization and Computer Graphics, IEEE

Transactions on, 14(6):1277–1284.

Debiasi, A., Sim

oes, B., and De Amicis, R. (2014). Super-

vised force directed algorithm for the generation of

ﬂow maps. In Proceedings of the WSCG 2014 - 22nd

International Conference on Computer Graphics.

Debiasi, A., Sim

oes, B., and De Amicis, R. (2015a).

Schematization of node-link diagrams and drawing

techniques for geo-referenced networks. In Cy-

berworlds (CW), 2015 International Conference on.

IEEE.

Debiasi, A., Simes, B., and Amicis, R. D. (2015b). 3dar-

clens: Interactive network analysis on geographic sur-

faces. In Proceedings of the 6th International Confer-

ence on Information Visualization Theory and Appli-

cations, pages 291–299.

Ellis, G. and Dix, A. (2007). A taxonomy of clutter

reduction for information visualisation. Visualiza-

tion and Computer Graphics, IEEE Transactions on,

13(6):1216–1223.

Ersoy, O., Hurter, C., Paulovich, F. V., Cantareiro, G.,

and Telea, A. (2011). Skeleton-based edge bundling

for graph visualization. Visualization and Computer

Graphics, IEEE Transactions on, 17(12):2364–2373.

Gansner, E. R., Hu, Y., North, S., and Scheidegger, C.

(2011). Multilevel agglomerative edge bundling for

visualizing large graphs. In Paciﬁc Visualization

Symposium (PaciﬁcVis), 2011 IEEE, pages 187–194.

IEEE.

Ghoniem, M., Fekete, J.-D., and Castagliola, P. (2004). A

comparison of the readability of graphs using node-

link and matrix-based representations. In Information

Visualization, 2004. INFOVIS 2004. IEEE Symposium

on, pages 17–24. Ieee.

Hadlak, S., Schumann, H., and Schulz, H.-J. (2015). A

survey of multi-faceted graph visualization. In Borgo,

R., Ganovelli, F., and Viola, I., editors, Eurographics

Conference on Visualization (EuroVis) - STARs. The

Eurographics Association.

Holten, D. (2006). Hierarchical edge bundles: Visualization

of adjacency relations in hierarchical data. Visualiza-

tion and Computer Graphics, IEEE Transactions on,

12(5):741–748.

Holten, D. and Van Wijk, J. J. (2009). Force-directed edge

bundling for graph visualization. In Computer Graph-

ics Forum, volume 28, pages 983–990. Wiley Online

Library.

Hurter, C., Ersoy, O., and Telea, A. (2012). Graph bundling

by kernel density estimation. In Computer Graphics

Forum, volume 31, pages 865–874. Wiley Online Li-

brary.

Hurter, C., Telea, A., and Ersoy, O. (2011). Moleview:

An attribute and structure-based semantic lens for

Schematization of Clutter Reduction Techniques in Geographic Node-link Diagrams using Task-based Criteria

115

large element-based plots. Visualization and Com-

puter Graphics, IEEE Transactions on, 17(12):2600–

2609.

Keim, D. A., Mansmann, F., Schneidewind, J., and Ziegler,

H. (2006). Challenges in visual data analysis. In In-

formation Visualization, 2006. IV 2006. Tenth Inter-

national Conference on, pages 9–16. IEEE.

Lambert, A., Bourqui, R., and Auber, D. (2010a). 3d edge

bundling for geographical data visualization. In In-

formation Visualisation (IV), 2010 14th International

Conference, pages 329–335. IEEE.

Lambert, A., Bourqui, R., and Auber, D. (2010b). Wind-

ing roads: Routing edges into bundles. In Computer

Graphics Forum, volume 29, pages 853–862. Wiley

Online Library.

Lee, B., Plaisant, C., Parr, C. S., Fekete, J.-D., and Henry,

N. (2006). Task taxonomy for graph visualization.

In Proceedings of the 2006 AVI workshop on BEyond

time and errors: novel evaluation methods for infor-

mation visualization, pages 1–5. ACM.

Liu, S., Cui, W., Wu, Y., and Liu, M. (2014). A survey on

information visualization: recent advances and chal-

lenges. The Visual Computer, 30(12):1373–1393.

Luo, S.-J., Liu, C.-L., Chen, B.-Y., and Ma, K.-L. (2012).

Ambiguity-free edge-bundling for interactive graph

visualization. Visualization and Computer Graphics,

IEEE Transactions on, 18(5):810–821.

Munzner, T., Hoffman, E., Claffy, K., and Fenner, B. G.

N. D. (1996). Visualizing the global topology of the

mbone. In Information Visualization’96, Proceedings

IEEE Symposium on, pages 85–92. IEEE.

Nocaj, A. and Brandes, U. (2013). Stub bundling and

conﬂuent spirals for geographic networks. In Graph

Drawing, pages 388–399. Springer.

Panagiotidis, A., Bosch, H., Koch, S., and Ertl, T. (2011).

Edgeanalyzer: Exploratory analysis through advanced

edge interaction. In System Sciences (HICSS), 2011

44th Hawaii International Conference on, pages 1–

10. IEEE.

Peng, D., Lu, N., Chen, W., and Peng, Q. (2012). Side-

knot: Revealing relation patterns for graph visualiza-

tion. In Paciﬁc Visualization Symposium (PaciﬁcVis),

2012 IEEE, pages 65–72. IEEE.

Peysakhovich, V., Hurter, C., and Telea, A. (2015).

Attribute-driven edge bundling for general graphs

with applications in trail analysis. In 2015 IEEE

Paciﬁc Visualization Symposium, PaciﬁcVis 2015,

Hangzhou, China, April 14-17, 2015, pages 39–46.

Phan, D., Xiao, L., Yeh, R., and Hanrahan, P. (2005). Flow

map layout. In Information Visualization, 2005. IN-

FOVIS 2005. IEEE Symposium on, pages 219–224.

IEEE.

Pupyrev, S., Nachmanson, L., Bereg, S., and Holroyd, A. E.

(2012). Edge routing with ordered bundles. In Graph

Drawing, pages 136–147. Springer.

Purchase, H. C., Carrington, D., and Allder, J. (2004). Eval-

uating graph drawing aesthetics: deﬁning and explor-

ing a new empirical research area. Computer Graphics

and Multimedia, pages 145–178.

Qu, H., Zhou, H., and Wu, Y. (2007). Controllable and pro-

gressive edge clustering for large networks. In Graph

Drawing, pages 399–404. Springer.

Riche, N. H., Dwyer, T., Lee, B., and Carpendale, S. (2012).

Exploring the design space of interactive link curva-

ture in network diagrams. In Proceedings of the In-

ternational Working Conference on Advanced Visual

Interfaces, pages 506–513. ACM.

Rosenholtz, R., Li, Y., Mansﬁeld, J., and Jin, Z. (2005).

Feature congestion: a measure of display clutter. In

Proceedings of the SIGCHI conference on Human fac-

tors in computing systems, pages 761–770. ACM.

Schmidt, S., Nacenta, M. A., Dachselt, R., and Carpendale,

S. (2010). A set of multi-touch graph interaction tech-

niques. In ACM International Conference on Interac-

tive Tabletops and Surfaces, pages 113–116. ACM.

Selassie, D., Heller, B., and Heer, J. (2011). Divided

edge bundling for directional network data. Visualiza-

tion and Computer Graphics, IEEE Transactions on,

17(12):2354–2363.

Speckmann, B. and Verbeek, K. (2010). Necklace maps.

IEEE Trans. Vis. Comput. Graph., 16(6):881–889.

Sun, G.-D., Wu, Y.-C., Liang, R.-H., and Liu, S.-X. (2013).

A survey of visual analytics techniques and appli-

cations: State-of-the-art research and future chal-

lenges. Journal of Computer Science and Technology,

28(5):852–867.

Telea, A. and Ersoy, O. (2010). Image-based edge bundles:

Simpliﬁed visualization of large graphs. In Computer

Graphics Forum, volume 29, pages 843–852. Wiley

Online Library.

Tominski, C., Gladisch, S., Kister, U., Dachselt, R., and

Schumann, H. (2014). A survey on interactive lenses

in visualization. EuroVis State-of-the-Art Reports,

pages 43–62.

Verbeek, K., Buchin, K., and Speckmann, B. (2011). Flow

map layout via spiral trees. Visualization and Com-

puter Graphics, IEEE Transactions on, 17(12):2536.

Wong, N. and Carpendale, S. (2007). Supporting interactive

graph exploration with edge plucking. Proc. VDA’07.

Wong, N., Carpendale, S., and Greenberg, S. (2003). Edge-

lens: An interactive method for managing edge con-

gestion in graphs. In Information Visualization, 2003.

INFOVIS 2003. IEEE Symposium on, pages 51–58.

IEEE.

Xu, K., Rooney, C., Passmore, P., Ham, D.-H., and Nguyen,

P. H. (2012). A user study on curved edges in graph

visualization. Visualization and Computer Graphics,

IEEE Transactions on, 18(12):2449–2456.

Zhou, H., Xu, P., Yuan, X., and Qu, H. (2013). Edge

bundling in information visualization. Tsinghua Sci-

ence and Technology, 18(2):145–156.

Zhou, H., Yuan, X., Cui, W., Qu, H., and Chen, B.

(2008). Energy-based hierarchical edge clustering

of graphs. In Visualization Symposium, 2008. Paci-

ﬁcVIS’08. IEEE Paciﬁc, pages 55–61. IEEE.

IVAPP 2016 - International Conference on Information Visualization Theory and Applications

116