Visualization of Large Networks1

VLADIMIR BATAGELJ

University of Ljubljana, Ljubljana, Slovenia

Article Outline

GlossaryI. DefinitionII. IntroductionIII. AttemptsIV. PerspectivesBibliography

Glossary

For the basic notions on graphs and networks see the articles Wouter de Nooy:Social network analysis and Vladimir Batagelj: Social Network Analysis, Large-scale in the Social Networks section. For a complementary information on graphdrawing in social network analysis see the article Linton Freeman: Methods ofSocial Network Visualisation.

k-core – a set of vertices in a graph is a k-core if each vertex from the set hasinternal (restricted to the set) degree at least k and the set is maximal – no suchvertex can be added to it.

Network – consists of vertices linked by lines and additional data about verticesand/or lines. A network is large if it has at least some hundreds vertices and canbe stored in a computer memory.

Partition – partition of a set is a family of its nonempty subsets such that eachelement of the set belongs to exactly one among the subsets. The subsets are alsocalled classes or groups.

1To be published as a chapter in the Encyclopedia of Complexity and System Science (editor-in-chief Bob Meyers), in the Complex Networks section (section editor Geoffrey Canright),Springer Verlag, 2009.version: May 8, 2008 / 00 : 44

1

Spring embedder – is another name for energy minimization graph drawingmethod. The vertices are considered as particles with repulsive force among them,and lines as springs that attract/repel the vertices if they are too far/close. Thealgorithm is trying to determine an embedding of vertices in two or three dimen-sional space that minimizes the ’energy’ of this system.

I. Definition

The earliest pictures containing graphs were magic figures, connections betweendifferent concepts (for example the Sephirot in Jewish Kabbalah), game boards(nine men’s morris, pachisi, patolli, go, xiangqi, and others) road maps (for ex-ample Roman roads in Tabula Peutingeriana), and genealogical trees of importantfamilies [29].

The notion of graph was introduced by Euler. In the 18th and 19th centurythe graphs were mainly used for solving different recreational problems (Knight’stour, Eulerian and Hamiltonian problem, map coloring). At the end of 19th cen-tury some applications of graphs to real life problems (electric circuits, Kirchhoff;molecular graphs, Kekule) appeared. In 20th century the graph theory evolvedinto a field of discrete mathematics of its own with applications to transportationnetworks (road and railway systems, metro lines, bus lines), project diagrams,flowcharts of computer programs, electronic circuits, molecular graphs, etc.

In social science the use of graphs was introduced by Jacob Moreno (1934)as a basis of his sociometric approach. In his book Who shall survive? [54] arelatively large network Sociometric geography of community – map III (435 in-dividuals, 4350 lines) is presented. Linton Freeman wrote a detailed account ofdevelopment of social network analysis [50] and about the visualization of socialnetworks [17]. The networks studied in the social network analysis till ninetieswere mostly small – some tens of vertices.

II. Introduction

In eighties the development of IT made a new field of computer graphics possibleto emerge. Soon also the first algorithms for graph drawing appeared:

• trees: Wetherell and Shannon, 1979 [40].

• acyclic graphs: Sugiyama, 1981 [37];

• energy minimization methods (spring embedders) for general graphs: Eades1984 [15], Kamada and Kawai, 1989 [27], Fruchterman and Reingold, 1991 [18];

2

18 Aug 05 -- 1 of 19 -- CV_Klovdahl_180805USL02a.doc

CURRICULUM VITAE

Alden S. Klovdahl, Ph.D.

Image: The core (n à 450) of a social network ofover 5,000 urban residents

(in Canberra, Australia).

Copyright 1995, 2005 A.S. Klovdahl

All Rights Reserved.

[ klovdahl.gif ]

alden.klovdahl@anu.edu.au

Figure 1: Klovdahl: social links in Canberra, Australia

In energy minimization methods vertices are considered as particles with repulsiveforce among them and lines as springs that attract/repel the vertices if they are toofar/close. The algorithms are trying to determine an embedding of vertices in twoor three dimensional space that minimizes the ’energy’ of this system.

Already in 1963 William Tutte proposed an algorithm for drawing planargraphs [38] and Donald E. Knuth an algorithm for drawing flowcharts [19].

A well known example of an early graph visualization was produced by AldenKlovdahl using his program View Net – see Figure 1. Nice as a piece of art, butwith an important message: there are big problems with visualization of denseor/and large graphs.

These developments led to a new field of graph drawing to emerge. In 1992a group of computer scientists and mathematicians (Giuseppe Di Battista, PeterEades, Michael Kaufmann, Pierre Rosenstiehl, Kozo Sugiyama, Roberto Tamas-sia, Ioannis Tollis, and others) started the conference International Symposium onGraph Drawing which takes place each year. The proceedings of the conferenceare published in the Lecture Notes in Computer Science series by Springer [51].To stimulate new approaches to graph drawing a graph drawing contest is acom-panying each conference. Many papers on graph drawing are published in theJournal of Graph Algorithms and Applications [88].

3

Figure 2: Network of traceroute paths for 29 June 1999

Most of the efforts of the graph drawing community were spent on problemsof drawing of special types of graphs (trees, acyclic graphs, planar graphs) orusing special styles (straight lines, orthogonal, grid-based, circular, hierarchical)and deriving bounds on required area for selected type of drawing.

In nineties further development of IT (GUI, multimedia, WWW) made largegraphs analysis a reality. For example studies of: large organic molecules (PDB [60]),Internet (Caida [61]), and genealogies (White and Jorion [41], FamilySearch [69]).In chemistry several tools for dynamic 3D visualization and inspection of moleculeswere developed (Kinemage, 1992 [90], Rasmol, 1993 [97], MDL Chime, 1996 [91]).

One of the earliest systems for large networks was SemNet (Fairchild, Poltrock,Furnas, 1988 [16]) used to explore knowledge bases represented as directed graphs.

In 1991 the Tom Sawyer Software [101] was founded – the premier providerof high performance graph visualization, layout, and analysis systems that enablethe user to see and interpret complex information to make better decisions.

In 1993 at AT&T started the development of GraphViz tools for graph visual-ization (dot, neato, dotty, tcldot, libgraph) [78]. Becker, Eick and Wilks developeda SeeXYZ family of network visualization programs [8].

In 1996 Vladimir Batagelj and Andrej Mrvar started the development of Pajek–

4

a program for large network analysis and visualization [5].In 1997 at La Sepienza, Rome started the development of GDToolkit [73] as

an extension of LEDA (Library of Efficient Data types and Algorithms) to provideimplementations of several classical graph-drawing algorithms. The new versionGDT 4.0 (2007) produced in collaboration with University of Perugia is LEDAindependent.

Graham Wills a principal invetigator at Bell Labs (1992-2001) built the Nicheworkssystem for the visual analysis of very large (up to million vertices) weighted net-work graphs.

In the summer of 1998 Bill Cheswick and Hal Burch started to work on theInternet Mapping Project at Bell Labs [62]. Its goal was to acquire and saveInternet topological data over a long period of time. This data has been usedin the study of routing problems and changes, DDoS attacks, and graph theory.In the fall of 2000 Cheswick and Burch moved to a spin-off from Lucent/BellLabs named Lumeta Corporation. Bill Cheswick is now back at AT&T Labs. InFigure 2 a network obtained from traceroute paths for 29 June 1999 with nearly100000 vertices is presented.

In the years 1997-2004 Martin Dodge maintained his Cybergeography Re-search web pages [67]. The results were published in the book The Atlas ofCyberspace [48]. A newer, very rich site on information visualization is Visualcomplexity [105] where many interesting ideas on network visualizations can befound. These examples, and many others, can be accessed also from the Infovis1100+ examples of information visualization site [86]. An interesting collectionof graph/network visualizations can be found also on CDs of Gerhard Dirmoser.The collection contains also many artistic and other pictures not produced by com-puters.

In 1997 Harald Katzmair founded FAS research, Vienna, Austria [70] a com-pany providing network analysis services. They emphasize the importance ofnice looking final products (pictures) for costumers by enhancing, using graphi-cal tools, the visual quality of the results obtained by network analysis tools. InFigure 3 a network of Austrian research projects is presented. Similar companyAguidel [58] was founded in France by Andrei Mogoutov, author of the programReseau-Lu.

Every year (from 2002) at the INSNA Sunbelt conference [87] the Viszardsgroup has a special session in which they present their solutions – analyses andvisualizations of selected network or type of networks. Most of the selected net-works were large (KEDS, Internet Movie Data Base, Wikipedia, Web of Science).

5

Figure 3: FAS: The scientific field of Austria

6

Figure 4: CAIDA: AS core 2008

III. Attempts

In the new millenium several attempts to develop programs for drawing largegraphs/networks were made. Most of the following descriptions are taken verba-tim from the programs’ web pages.

Stephen Kobourov with his collaborators from University of Arizona devel-oped two graph drawing systems GRIP (2000) [20, 79] and Graphael (2003) [75].GRIP – Graph dRawing with Intelligent Placement was designed for drawinglarge graphs and uses a multi-dimensional force-directed method together withfast energy function minimization. It employs a simple recursive coarseningscheme – rather than being placed at random, vertices are placed intelligently,several at a time, at locations close to their final positions.

The supercomputer centers Cooperative Association for Internet Data Anal-ysis (CAIDA) [61], co-founded in 1998 by kc claffy, is an independent researchgroup dedicated to investigating both the practical and theoretical aspects of theInternet to promote the engineering and maintenance of a robust, scalable globalInternet infrastructure. They have primarily been focusing on understanding howthe Internet is evolving, and developing a state-of-the-art infrastructure for datameasurement that can be shared with the entire research community.

Figure 4 represents a macroscopic snapshot of the Internet for two weeks:1-17 January 2008. The graph reflects 4853991 observed IPv4 addresses and

7

Figure 5: Walrus

5682419 IP links. The network is aggregated into a topology of Autonomous Sys-tems(ASes). The abstracted graph consists of 17791 ASes (vertices) and 50333peering sessions (lines).

Walrus is a tool for interactively visualizing large directed graphs in three-dimensional space. It is best suited to visualizing moderately sized (a few hun-dred thousand vertices) graphs that are nearly trees. Walrus uses 3D hyperbolicgeometry to display graphs under a fisheye-like distortion. By bringing differentparts of a graph to the magnified central region, the user can examine every part ofthe graph in detail. Walrus was developed by Young Hyun at CAIDA based on re-search by Tamara Munzner. In Figure 5 two examples of visualizations producedwith Walrus are presented.

Some promissing algorithms for drawing large graphs were proposed by Ul-rik Brandes, Tim Dwyer, Emden Gansner, Stefan Hachul, David Harel, MichaelJunger, Yehuda Koren, Andreas Noack, Stephen North, Christian Pich, and ChrisWalshaw [39, 28, 21, 23, 14, 34, 22, 12]. They are based either on multilevelenergy minimization approach or on algebraic or spectral approach that reducesto some application of eigenvectors.

Multilevel approach speeds-up the algorithms. Multilevel algorithms are basedon two phases: a coarsening phase, in which a sequence of coarse graphs with de-creasing sizes is computed and a refinement phase in which successively drawingsof finer graphs are computed, using the drawings of the next coarser graphs anda variant of a suitable force-directed single-level algorithm [22]. The fastest al-gorithms combine the multilevel approach with fast approximation of long rangerepulsive force using nested data structure, such as quadtree or kd-tree.

8

Figure 6: Katy Borner: Text analysis

9

Katy Borner from Indiana University with her collaborators produced severalvisualizations of diferent scientometric networks such as Backbone of Science [9]and Wikipedia [82]. They are using different visual cues to produce informationrich visualizations – see Figure 6. She also commissioned the Map of Sciencebased on data (800000 published papers) from Thomson ISI and produced byKevin Boyack, Richard Klavans and Bradford Paley [9].

Yifan Hu from AT&T Labs Information Visualization Group developed a mul-tilevel graph drawing algorithm for visualization of large graphs [26]. The algo-rithm was first implemented in 2004 in Mathematica and released in 2005. Fordemonstration he applied it to the University of Florida Sparse Matrix collec-tion [65] that contains over 1500 square matrices. The results are available inthe Gallery of Large Graphs [83]. The largest graph (vanHeukelum/cage15) has5154859 vertices and 47022346 edges. In Figure 7 selected pictures from theGallery are presented.

From the examples that we have given, we can see that, in some cases, thegraph drawing algorithms can detect symmetries in a given graph and also a’structure’ ((sub)trees, clusters, planarity, etc.). The main problem are graphs withdense part(s).

For dense parts a better approach is to display them using matrix represen-tation. This representation was used in 1999 by Vladimir Batagelj, Andrej Mr-var and Matjaz Zaversnik in their partitioning approach to visualization of largegraphs [6] and is a basis of systems Matrix Zoom by James Abello and Frank vanHam, 2004 [1, 2], and MatrixExplorer by Nathalie Henry and Jean-Daniel Fekete,2006 [25]. A matrix representation is determined by an ordering of vertices. Thereexist several algorithms to produce the orderings. A comparative study of themwas published by Chris Mueller [32, 93]. In Figure 8 three orderings of the samematrix are presented. Most ordering algoritms were designed for applications innumerical, and not data, analysis. The orderings can be also determined usingclustering or blockmodeling methods [49].

An important type of networks are temporal networks where presence of ver-tices and/or lines changes through time. Visualization of such networks requiresspecial approaches (Sonia [98], SVGanim [100], TecFlow [84]). An interestingapproach to visualization of temporal networks was developed by Ulrik Brandesand his group [11].

IV. Perspectives

In this section a collection of ideas how to approach visualization of large net-works is presented. They are only partially implemented in different visualizationsolutions.

10

Figure 7: Examples from the Gallery of Large Graphs

11

Figure 8: Matrix representations

Figure 9: Big picture, V. Batagelj, AE’04

While the technical problems of graph drawing could ask for a single ’best’picture, the network analysis is also a part of data analysis. Its goal is to getinsight not only into the structure and characteristics of a given network, but alsointo how this structure influences processes going on over the network. We usuallyneed several pictures to present the obtained results.

Small graphs can be presented in their totality and in detail within a singleview. In a comprehensive view of large graphs, details become lost – converselya detailed view can encompass only a part of the graph.

The literature on graph drawing, is dominated by the ’sheet of paper’ paradigm

12

– the solutions and techniques are mainly based on the assumption that the finalresult is a static picture on a sheet of paper. In this case to present a large data setwe need a large ’sheet of paper’ – but this has a limit. In Figure 9 a visualization ispresented of the symmetrized subnetwork of 5952 words and 18008 associationsfrom the Edinburgh Associative Thesaurus [68] prepared by Vladimir Batagelj on3m × 5m ’sheet of paper’ for Ars Electronica, Linz 2004 Language of networksexhibition.

The main tool to deal with large objects is abstraction. In graphs it is usuallyrealized using a hierarchy of partitions. Shrinking selected classes of the partitionwe obtain a smaller reduced graph. The main operations related to abstraction are:

• cut-out: display only selected parts (classes of partition) of graph;

• context: detailed display of selected parts (classes) of graph and the displayof the rest in the reduced form;

• model: the reduced graph with respect to a given partition;

• hierarchy: the tree representing the nesting of graph partitions;

In larger/denser networks there is often too much information to be presentedat once. A possible answer are interactive layouts on computer screen where theuser controls what (s)he wants to see.

The computer screen is a different medium, which offers many new possibil-ities: parallel views (global and local); brushing and linking; zooming and pan-ning; temporary elements (additional information about the selected elements,labels, legends, markers, etc.); highlighted selections; and others. These featurescan and should be maximally leveraged to support data analytic tasks; or repeatingthe Shneiderman’s mantra: overview first, zoom and filter, then details on-demand(extended with: relate, history and extract) [35].

Interactively inspecting very large graphs a serious problem appears: How toavoid the “lost within the forest” effect? There are several solutions that can helpthe user maintain orientation:

• restart option: returns the user to the starting point;

• introduction of additional orientation elements. These elements can beswitched on and off.

• multiview: consists of at least two views (windows):

– map view: overall global view which contains the current position andallows ’long’ moves (jumps). For very large graphs it can be combinedwith zooming or fish-eye views.

13

Figure 10: Glasses: Rasmol displays – BallStick, SpaceFill, Backbone, Ribbons

– local view: displays a selected portion of the graph.

Additional support can be achieved by implementing trace/backtrack/replay mech-anism and guided tours.

In an interactive dynamic visualization of a graph on the computer screen itneeds not to be displayed in its totality. Inspecting a visualization the user can se-lect which parts and elements will be displayed and how. See for example Touch-Graph [102].

Closely related with the multiview concept are associated concepts of glasses,lenses and zooming. Glasses have effect on the entire window, and lenses only onthe selected region or elements.

14

Figure 11: Glasses: Display of properties – school

Selecting different glasses we obtain different views on the same data – sup-porting different visualisation aims. For example in Figure 10 four differentglasses (ball and stick, space-fill, backbone, ribbons) were applied in programRasmol to the molecula 1atn.pdb (deoxyribonuclease I complex with actin).

Another example of glasses is presented in Figure 11. The two pictures wereproduced by James Moody [31]. The graph picture was obtained by spring em-bedder. It represents the friendship relation among students in a school. Theglasses are the coloring of its vertices by different partitions: age partition (leftpicture) and race partition (right picture). This gives us the explanation of the fourgroups in the obtained graph picture – they are characterized by younger / olderand white / black students.

In Figure 12 a part of the big picture (Figure 9) is presented. The glasses in thiscase are based on ordering the edges in increasing order of their values and draw-ing them in this order – stronger edges covers the weaker. The picture emphasizesthe strongest substructures; the remaining elements form a background.

There are many kinds of glasses in representation of graphs. For example fish-eye views, matrix representation, using application field conventions (genealogies,molecules, electric circuits, SBGN), displaying vertices only, selecting the typeof labels (long/short name, value), displaying only the important vertices and/orlines, size of vertices determined by core number or betweenes.

An example of lens is presented in Figure 13 – contributions of companiesto different presidential candidates from Follow the Oil Money by Greg Michalecand Skye Bender-deMoll [71]. For a selected vertex the information about it isdisplayed. Another example of lens would be to temporarily enhance the displayof neighbors of the selected vertex [104] or to display their labels. The ”shak-ing” option used in Pajek to visually identify all vertices from selected cluster

15

Figure 12: Part of the big picture

Figure 13: Lenses: temporary info about the selected vertex

is also a kind of lense; so are the matrix representations of selected clusters inNodeTrix [81].

Additional enhancement of a presentation can be achieved by the use of sup-port elements such as labels, grids, legends, and various forms of help facilities.

16

Figure 14: Zoom, glasses, lenses, navigation: Google Maps

An important concept connected with zooming is the LOD (Level of Detail) –subobjects are displayed differently depending on the zoom depth.

A nice example of combination of these techniques is the Google Maps ser-vice [74] – see Figure 14. It combines zooming, glasses (Map, Satellite, Terrain),navigation (left, right, up, down) and lenses (info about points). The maps at dif-ferent zoom levels provide information at different level of detail and in differentform.

For a given hierarchical clustering of its vertices a similar approach could beused also for inspection of large graphs/networks. To produce higher level ’maps’different methods can be used: k-core representation [4], density contours [92],generalized blockmodeling [49], clustering [80] (Figure 15), preserving only im-portant vertices and lines, etc. In visualizing the ’maps’ new graphical elements(many of them still to be invented) can be used [13, 89][49, p. 223] preserv-ing/indicating the information about the structure at lower level.

The k-core representation [4, 94] is based on k-core decomposition of a net-work [6, 7] and was developed by Alessandro Vespignani and his collaborators [63].In Figure 16 a portion of the web at .fr domain, with 1 million of pages is pre-sented. Each node represents a web page and each edge is a hyperlink betweentwo pages.

Density contours were introduced by James Moody in 2006. First a springembedder layout of (valued) network is determined. Afterward the vertices andlines are removed and replaced by density contours. In Figure 17 this process ispresented in the case of social science co-citation network. On the left side the net-work layout is presented and in the right bottom part we can see the corresponding

17

Figure 15: Zoom, glasses, lenses, navigation: Grokker

Figure 16: k-core structure of a portion of the web at .fr domain

18

Figure 17: Density structure

density contours.The basic steps in graph/network visualization are:

graph/network→ analyses → layouts→ viewer → pictures

On this scheme the development of different tools can be based depending on thekind of users (simple, advanced) and their tasks (reporting, learning, monitoring,exploration, analysis) they address. In some cases only a viewer will be sufficient(for example SVG viewer, X3D viewer, or a special graph layout viewer), in othersa complete network analysis system is needed (such as Geomi [3, 72], ILOG [85],Pajek [59], Tulip [103], yFiles [106]).

Layouts are obtained by augmenting the network data with results of analysesand user’s decisions to be used to visualize the network. In Pajek’s input for-mat there are several layout elements from Pajek’s predecessors (see Pajek’smanual, pages 69-73). As in typesetting

text + formatting = formated textso in network visualization

network + layout = pictureIt would be useful to define a common layout format (an extension of GraphML [77]

?) so that independent viewer modules can be developed and combined with dif-ferent layout algorithms. Some useful ideas can be found in nViZn (”envision”)system [99]. To specify layouts we can borrow from the typesetting the notion ofstyle.

19

Bibliography

Primary Literature

[1] Abello J., van Ham F. (2004) Matrix Zoom: A Visual Interface to Semi-external Graphs. IEEE Symposium on Information Visualization 2004, Oc-tober 10-12, Austin, Texas, USA, 183-190.

[2] Abello J., van Ham F., Neeraj Krishnan (2006) ASK-GraphView : A LargeScale Graph Visualization System. IEEE Transactions on Visualization andComputer Graphics, Vol. 12, No. 5, September/October 2006.

[3] Ahmed A., Dwyer T., Forster M., Fu X., Ho J., Hong S., Koschutzki D.,Murray C., Nikolov N., Taib R., Tarassov A., Xu K. (2006). GEOMI: GE-Ometry for Maximum Insight. In Healy P., Eades P. (eds.), Proc. 13th Int.Symp. Graph Drawing (GD2005), Springer-Verlag, Lecture Notes in Com-puter Science, Vol. 3843, pp. 468-479.

[4] Alvarez-Hamelin J.I., DallAsta L., Barrat A., Vespignani A. (2005). Largescale networks fingerprinting and visualization using the k-core decomposi-tion. Advances in Neural Information Processing Systems 18 [Neural Infor-mation Processing Systems, NIPS 2005, December 5-8, 2005, Vancouver,British Columbia, Canada.

[5] Batagelj V., Mrvar A. (2003). Pajek – Analysis and Visualization of LargeNetworks. in Junger, M., Mutzel, P., (Eds.) Graph Drawing Software.Springer, Berlin, p. 77-103.

[6] Batagelj, V., Mrvar, A. and Zaversnik, M. (1999). Partitioning approach tovisualization of large graphs, In Kratochvıl, J. (ed), Lecture notes in com-puter science, 1731, Springer, Berlin, 90–97.

[7] Batagelj V., Zaversnik, M. (2002) Generalized Cores, arxiv cs.DS/0202039

[8] Becker R.A., Eick S.G., Wilks A.R. (1995). Visualizing network data. IEEETransactions on Visualization and Computer Graphics, 1(1):16-28.

20

[9] Boyack K.W., Klavans R., Brner K. (2005) Mapping the Backbone of Sci-ence. Scientometrics, 64 (3), 351-374.

[10] Boyack K.W., Klavans R., Paley W.B. (2006). Map of Science. Nature, Vol444, Dec. 2006.

[11] Brandes U., Fleischer D., Lerner J. (2006). Summarizing Dynamic Bipo-lar Conflict Structures. IEEE Transactions on Visualization and ComputerGraphics, special issue on Visual Analytics, 12(6):1486-1499.

[12] Brandes U., Pich C. (2007). Eigensolver Methods for Progressive Multidi-mensional Scaling of Large Data. Proc. 14th Intl. Symp. Graph Drawing(GD ’06). LNCS 4372, Springer-Verlag, pp. 42-53.

[13] Dickerson M., Eppstein D., Goodrich M.T., Meng J. (2005). Confluent draw-ings: visualizing non-planar diagrams in a planar way. J. Graph Algorithmsand Applications (special issue for GD’03) 9(1):31-52.

[14] Dwyer T., Koren Y. (2005). DIG-COLA: Directed Graph Layout throughConstrained Energy Minimization. INFOVIS 2005: 9

[15] Eades P. (1984). A heuristic for graph drawing. Congressus Numerantium,42:149-160.

[16] Fairchild K.M., Poltrock S.E., Furnas G.W. (1988). SemNet: Three-Dimensional Representations of Large Knowledge Bases. In R. Guindon(Ed.), Cognitive science and its applications for human-computer interac-tion. Hillsdale, New Jersey: Lawrence Erlbaum Associates, pp. 201233.

[17] Freeman L.C. (2000). Visualizing Social Networks. Journal of Social Struc-ture, Volume 1(1).

[18] Fruchterman T., Reingold E. (1991). Graph Drawing by Force DirectedPlacement. Software Practise and Experience, 21(11).

[19] Knuth D.E. (1963). Computer-drawn flowcharts. Communications of theACM, Vol. 6, 9, 555 - 563.

[20] Gajer P., Kobourov S. (2001). GRIP: Graph Drawing with Intelligent Place-ment. In Graph Drawing 2000, volume 1984 of LNCS, pages 222228.Springer-Verlag.

[21] Gansner E.R., Koren Y., North S.C. (2005). Topological Fisheye Views forVisualizing Large Graphs. IEEE Trans. Vis. Comput. Graph. 11(4): 457-468.

21

[22] Hachul S., Junger M. (2007) Large-Graph Layout Algorithms at Work: AnExperimental Study. JGAA, 11(2) 345369.

[23] Harel D., Koren Y. (2004). Graph Drawing by High-Dimensional Embed-ding. Journal of Graph Algorithms and Applications, vol. 8, no. 2, pp.195214.

[24] Herman I., Melancon G., Marshall M.S. (2000). Graph visualization andnavigation in information visualization: A survey. IEEE Transactions on Vi-sualization and Computer Graphics, 6(1), 24-43.

[25] Henry N., Fekete J-D. (2006). MatrixExplorer: a Dual-Representation Sys-tem to Explore Social Networks. IEEE Transactions on Visualization andComputer Graphics, Vol. 12, No. 5, September/October 2006.

[26] Hu Y.F. (2005). Efficient and High Quality Force-Directed Graph Drawing.Mathematica Journal, Vol. 10, 37-71.

[27] Kamada T., Kawai S. (1988). An Algorithm for Drawing General UndirectedGraphs. Information Processing Letters, 31:7-15, 1988.

[28] Koren Y. (2003). On Spectral Graph Drawing. COCOON 2003, 496-508.

[29] Kruja E., Marks J., Blair A., Waters R. (2001). A Short Note on the History ofGraph Drawing. Proc. of Graph Drawing 2001 (Springer LNCS, Vol. 2265),pp. 272-286.

[30] Lamping J., Rao R., Pirolli P. (1995). A Focus+Context Technique Based onHyperbolic Geometry for Visualizing Large Hierarchies. CHI 95: 401-408.

[31] Moody J. (2001). Race, School Integration, and Friendship Segregation inAmerica. American Journal of Sociology 107(3) 679:716.

[32] Mueller C., Martin B., Lumsdaine A. (2007). A Comparison of Vertex Or-dering Algorithms for Large Graph Visualization. APVIS 2007.

[33] Munzner T. (1997). H3: Laying Out Large Directed Graphs in 3D Hyper-bolic Space. Proceedings of the 1997 IEEE Symposium on Information Vi-sualization, October 20-21 1997, Phoenix, AZ, pp 2-10.

[34] Noack A. (2007). Energy Models for Graph Clustering. Journal of GraphAlgorithms and Applications 11(2):453-480, .

[35] Shneiderman B. (1996). The Eyes Have It: A Task by Data Type Taxon-omy for Information Visualization. In IEEE Conference on Visual Languages(VL’96). Boulder, CO: IEEE CS Press.

22

[36] Shneiderman B., Aris A. (2006) Network Visualization by Semantic Sub-strates. IEEE Trans. on Visualization and Computer Graphics, Vol. 12, No.5, pp. 733-740.

[37] Sugiyama K., Tagawa S., Toda M. (1981). Methods for Visual Understand-ing of Hierarchical Systems. IEEE Transactions on Systems, Man, and Cy-bernetics.

[38] Tutte W.T. (1963). How to Draw a Graph. Proceedings of the London Math-ematical Society 1963 s3-13(1):743-767.

[39] Walshaw C. (2003). A Multilevel Algorithm for Force-Directed Graph-Drawing. Journal of Graph Algorithms and Applications, vol. 7, no. 3, pp.253285

[40] Wetherell C., Shannon A. (1979). Tidy drawing of trees. IEEE Trans. Soft-ware Engineering, 5:514-520.

[41] White D.R., Jorion P. (1992). Representing and Computing Kinship: A NewApproach. Current Anthropology, Vol. 33, No. 4., pp. 454-463.

[42] Wills G.J. (1999). NicheWorks-Interactive Visualization of Very LargeGraphs. Journal of Computational and Graphical Statistics, vol.8(2), pp 190-212.

Books and Reviews

[43] Bertin J. (1967). Semiologie Graphique. Les diagrammes, les reseaux, lescartes. Mouton, Gauthier-Villars, Paris, La Haye.

[44] Brandes U., Erlebach T. (Eds.) (2005). Network Analysis: MethodologicalFoundations. LNCS, Springer, Berlin.

[45] Carrington P.J., Scott J., Wasserman S. (Eds.) (2005). Models and Methodsin Social Network Analysis. Cambridge University Press.

[46] de Nooy W., Mrvar A., Batagelj V. (2005) Exploratory Social Network Anal-ysis with Pajek, Cambridge University Press.

[47] di Battista G., Eades P., Tamassia R., Tollis I.G.(1999) Graph Drawing: Al-gorithms for the Visualization of Graphs. Prentice Hall.

23

[48] Dodge M., Kitchin R. (2002). The Atlas of Cyberspace. Pearson Education.

[49] Doreian P., Batagelj V., Ferligoj A. (2005) Generalized Blockmodeling,Cambridge University Press.

[50] Freeman L.C. (2004) The Development of Social Network Analysis: A Studyin the Sociology of Science. Empirical Press.

[51] Graph Drawing. Lecture Notes in Computer Science Vol. 894 (1994), 1027(1995), 1190 (1996), 1353 (1997), 1547 (1998), 1731 (1999), 1984 (2000),2265 (2001), 2528 (2002), 2912 (2003), 3383 (2004), 3843 (2005), 4372(2006), 4875 (2007). Springer, Berlin.

[52] Junger M., Mutzel P., (Eds.)( 2003). Graph Drawing Software. Springer,Berlin.

[53] Kaufmann M., Wagner D. (2001). Drawing Graphs, Methods and Models.Springer, Berlin.

[54] Moreno J.L. (1953). Who shall survive? Beacon House Inc., New York.

[55] Tufte E.R. (1983). The Visual Display of Quantitative Information. GraphicsPress, Cheshire.

[56] Wasserman S., Faust K. (1994). Social Network Analysis: Methods and Ap-plications. Cambridge University Press.

[57] Wilkinson L. (2000). The Grammar of Graphics. Statistics and Computing,Springer.

Web Resources

[58] Aguidel: http://www.aguidel.com/en/

[59] Batagelj V., Mrvar A.(1996-) Pajek– program for analysis and visualiza-tion of large network: http://pajek.imfm.sidata sets: http://vlado.fmf.uni-lj.si/pub/networks/data/

[60] Brookhaven Protein Data Bank: http://www.rcsb.org/pdb/

[61] Caida: http://www.caida.org/home/walrus gallery: http://www.caida.org/tools/visualization/walrus/gallery1/

24

[62] Cheswick B.: Internet Mapping Project – Map galleryhttp://www.cheswick.com/ches/map/gallery/

[63] Complex Networks Collaboratory: http://cxnets.googlepages.com/

[64] Cruz I., Tamassia R. (1994). Tutorial on graph drawing.http://graphdrawing.org/literature/gd-constraints.pdf

[65] Davis T.: University of Florida Sparse Matrix Collectionhttp://www.cise.ufl.edu/research/sparse/matrices

[66] Di Battista G., Eades P., Tamassia R., Tollis I.G. (1994). Algorithms forDrawing Graphs: an Annotated Bibliography. Computational Geometry:Theory and Applications 4:235-282, 1994.http://graphdrawing.org/literature/gdbiblio.pdf

[67] Dodge M.: Cyber-Geography Researchhttp://personalpages.manchester.ac.uk/staff/m.dodge/cybergeography/

[68] Edinburgh Associative Thesaurus (EAT): http://www.eat.rl.ac.uk/

[69] FamilySearch: http://www.familysearch.org/

[70] FAS.research, Vienna, Austria: http://www.fas.at/

[71] Follow the Oil Money: http://oilmoney.priceofoil.org/

[72] GEOMI (Geometry for Maximum Insight):http://www.cs.usyd.edu.au/˜visual/valacon/geomi/

[73] GDToolkit – Graph Drawing Toolkit:http://www.dia.uniroma3.it/˜gdt/gdt4/index.php

[74] Google Maps: http://maps.google.com/

[75] Graphael: http://graphael.cs.arizona.edu/

[76] Graphdrawing home page: http://graphdrawing.org/

[77] GraphML File Format: http://graphml.graphdrawing.org/

[78] Graphviz: http://graphviz.org/

[79] GRIP: http://www.cs.arizona.edu/˜kobourov/GRIP/

[80] Grokker – Enterprise Search Management and Content Integrationhttp://www.grokker.com/

25

[81] Henry N., Fekete J-D., Mcguffin M. (2007). NodeTrix: Hybrid Representa-tion for Analyzing Social Networks.https://hal.inria.fr/inria-00144496

[82] Herr B.W., Holloway T., Borner K. (2007). Emergent Mosaic of WikipedianActivity.http://www.scimaps.org/dev/big thumb.php?map id=158

[83] Hu Y.F.: Gallery of Large Graphs.http://www.research.att.com/˜yifanhu/GALLERY/GRAPHS/index1.html

[84] iCKN: TeCFlow – A Temporal Communication Flow Visualizer for SocialNetwork Analysis http://www.ickn.org/

[85] ILOG Diagrams: http://www.ilog.com/

[86] Infovis – 1100+ examples of information visualizationhttp://www.infovis.info/index.php?cmd=search&words=graph&mode=normal

[87] INSNA – International Network for Social Network Analysis:http://www.insna.org/

[88] Journal of Graph Algorithms and Applications: http://jgaa.info/

[89] KartOO visual meta search engine: http://www.kartoo.com/

[90] 3D Macromolecule analysis and Kinemage home pagehttp://kinemage.biochem.duke.edu/

[91] MDL Chime: http://www.mdli.com/

[92] Moody J. (2007). The Network Structure of Sociological Production II.http://www.soc.duke.edu/˜jmoody77/presentations/soc Struc II.ppt

[93] Mueller C.: Matrix visualizations.http://www.osl.iu.edu/˜chemuell/data/ordering/sparse.html

[94] LaNet-vi – Large Network visualization toolhttp://xavier.informatics.indiana.edu/lanet-vi/

[95] OLIVE, On-line Library of Information Visualization Environments.http://otal.umd.edu/Olive/

[96] Pad++: Zoomable User Interfaces: Portal filtering and ’magic lenses’.http://www.cs.umd.edu/projects/hcil/pad++/tour/lenses.html

26

[97] RasMol Home Page: http://www.umass.edu/microbio/rasmol/index2.htm

[98] Sonia – Social Network Image Animator:http://www.stanford.edu/group/sonia/

[99] SPSS nViZn: http://www.spss.com/research/wilkinson/nViZn/nvizn.html

[100] SVGanim: http://vlado.fmf.uni-lj.si/pub/networks/pajek/SVGanim

[101] Tom Sawyer Software: http://www.tomsawyer.com/home/index.php

[102] TouchGraph: http://www.touchgraph.com/

[103] Tulip: http://www.labri.fr/perso/auber/projects/tulip/

[104] Viegas F.B., Wattenberg M. (2007). Many Eyeshttp://services.alphaworks.ibm.com/manyeyes/page/Network Diagram.html

[105] Visual complexity: http://www.visualcomplexity.com/vc/

[106] yWorks / yFiles: http://www.yworks.com/en/products yfiles about.htm

27