On visualization of (social) networks Networks Approaches - - PowerPoint PPT Presentation

on visualization of social networks
SMART_READER_LITE
LIVE PREVIEW

On visualization of (social) networks Networks Approaches - - PowerPoint PPT Presentation

Social networks visualization V. Batagelj On visualization of (social) networks Networks Approaches Statistics Drawing of Vladimir Batagelj large graphs Drawing of dense graphs IMFM Ljubljana and IAM UP Koper Network decomposition


slide-1
SLIDE 1

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

On visualization of (social) networks

Vladimir Batagelj

IMFM Ljubljana and IAM UP Koper

Big Data Visual Analytics Shonan meeting – November 8-11, 2015

  • V. Batagelj

Social networks visualization

slide-2
SLIDE 2

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Outline

1 Networks 2 Approaches 3 Statistics 4 Drawing of large graphs 5 Drawing of dense graphs 6 Network decomposition 7 Layouts 8 Problems 9 References

Vladimir Batagelj: vladimir.batagelj@fmf.uni-lj.si Current version of slides (November 8, 2015, 05 : 13): http://vlado.fmf.uni-lj.si/pub/slides/shonan.pdf

  • V. Batagelj

Social networks visualization

slide-3
SLIDE 3

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Networks

Alexandra Schuler/ Marion Laging-Glaser: Analyse von Snoopy Comics

A network N = (V, L, P, W) is based on two sets – a set of nodes (vertices) V, that repre- sent the selected units, and a set of links (lines) L, that rep- resent ties between units. They determine a graph. A line can be directed – an arc, or undi- rected – an edge; L = E ∪ A. Additional data about nodes or links can be known – the prop- erties (attributes) P of nodes and weights W of links.

Network = Graph + Data

  • V. Batagelj

Social networks visualization

slide-4
SLIDE 4

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

”Countryside” school district

James Moody (2001) AJS Vol 107, 3,679–716, friendship relation

Only small

  • r

sparse networks can be displayed readably. On large networks graph drawing algorithms can reveal their

  • veral

structure. Can we explain the

  • btained

structure?

  • V. Batagelj

Social networks visualization

slide-5
SLIDE 5

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Display of properties – school (Moody)

To understand networks we need (additional) data!

  • V. Batagelj

Social networks visualization

slide-6
SLIDE 6

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Approaches to large networks

From algorithmic complexity analysis: for dealing with large data (millions of units) we are limited to subquadratic algorithms. Most of large networks are sparse – Dunbar’s number. Approaches to large networks analysis and visualization:

  • statistics;
  • important parts:
  • decomposition;
  • identification of important elements and substructures.

Visualization: initial network exploration, reporting results, story telling. For additional details see [4, 3, 5].

  • V. Batagelj

Social networks visualization

slide-7
SLIDE 7

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

EAT all-degree distribution

  • 1

5 10 50 100 500 1 5 10 50 500 5000

EAT all−degree distribution

deg freq

  • V. Batagelj

Social networks visualization

slide-8
SLIDE 8

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

in/out-degree distributions

  • 1

5 10 50 100 500 1 100 10000

in−degree distribution

inDeg inFreq

  • 1

5 10 50 500 5000 1 2 5 10 20 50 100 200 500

  • ut−degree distribution
  • utDeg
  • utFreq

The in-degree distribution is ”scale-free”-like. The parameters can be determined using the package of Clauset, Shalizi and Newman. See also Stumpf, et al.: Critical Truths About Power Laws.

It is important to consider the direction of links!

  • V. Batagelj

Social networks visualization

slide-9
SLIDE 9

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

VOSviewer / journals

links → projection of nodes

There are many algorithms for drawing large graphs and networks [11] (Brandes and Pich – MDS [8]; VOSviewer - level of detail, zooming, contours [17]).

  • V. Batagelj

Social networks visualization

slide-10
SLIDE 10

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

VOSviewer

zooming

  • V. Batagelj

Social networks visualization

slide-11
SLIDE 11

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

VOSviewer

nodes → density of nodes → contours

  • V. Batagelj

Social networks visualization

slide-12
SLIDE 12

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Examples from the Gallery of Large Graphs

Yifan Hu

  • V. Batagelj

Social networks visualization

slide-13
SLIDE 13

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Dense networks

Yifan Hu developed a multilevel graph drawing algorithm for visualization

  • f large graphs. For demonstration he applied it to the University of

Florida Sparse Matrix collection. The results are available in the Gallery of Large Graphs [15]. From these examples we can see that, in some cases, the graph drawing algorithms can detect symmetries in a given graph and also a ’structure’ ((sub)trees, clusters, planarity, etc.). The main problem are graphs with dense part(s). For dense networks a better approach is to display them using matrix representation (Pajek [6], Matrix Zoom [1, 2], and MatrixExplorer [10]). A matrix representation is determined by an ordering of nodes. There exist several algorithms to produce the orderings [13, 12]. Most ordering algoritms were designed for applications in numerical, and not data, analysis. The orderings can be also determined using clustering or blockmodeling methods [9].

  • V. Batagelj

Social networks visualization

slide-14
SLIDE 14

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

A display of World Trade 1999 network

Afghanistan Albania Algeria Angola Argentina Armenia Australia Austria Azerbaijan Bahamas Bahrain Bangladesh Barbados Belarus Belgium-Lux Belize Benin Bermuda Bolivia Bosnia Herzg Brazil Bulgaria Burkina Faso Burundi Cambodia Cameroon Canada Cent.Afr.Rep Chad Chile China Colombia Congo Costa Rica Cote Divoire Croatia Cuba Cyprus Czech Rep Dem.Rp.Congo Denmark Djibouti Dominican Rp Ecuador Egypt El Salvador Eq.Guinea Estonia Ethiopia Falkland Is Fiji Finland Fr Ind O Fr.Guiana France,Monac Gabon Gambia Georgia Germany Ghana Gibraltar Greece Greenland Guadeloupe Guatemala Guinea GuineaBissau Guyana Haiti Honduras Hungary Iceland India Indonesia Iran Iraq Ireland Israel Italy Jamaica Japan Jordan Kazakhstan Kenya Kiribati Korea D P Rp Korea Rep. Kuwait Kyrgyzstan Lao P.Dem.R Latvia Lebanon Liberia Libya Lithuania Madagascar Malawi Malaysia Mali Malta Mauritania Mauritius Mexico Mongolia Morocco Mozambique Myanmar Nepal Neth.Ant.Aru Netherlands New Calednia New Zealand Nicaragua Niger Nigeria Norway Oman Oth.Oceania Pakistan Panama Papua N.Guin Paraguay Peru Philippines Poland Portugal Qatar Rep Moldova Romania Russian Fed Rwanda Samoa Saudi Arabia Senegal Seychelles Sierra Leone Singapore Slovakia Slovenia Somalia South Africa Spain Sri Lanka St.Helena St.Kt-Nev-An St.Pierre Mq Sudan Suriname Sweden Switz.Liecht Syria Taiwan Tajikistan Tanzania TFYR Macedna Thailand Togo Trinidad Tbg Tunisia Turkey Turkmenistan Uganda UK Ukraine Untd Arab Em Uruguay USA Uzbekistan Venezuela Viet Nam Yemen Yugoslavia Zambia Zimbabwe

  • V. Batagelj

Social networks visualization

slide-15
SLIDE 15

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Matrix display of World Trade 1999 network

Pajek - shadow [0.00,4.00] USA Japan Spain Netherland Belgium-Lu UK Germany Italy France,Mon Korea Rep. China India Thailand Indonesia Australia Singapore Malaysia South Afri Portugal Canada Brazil Mexico Argentina Turkey Greece Russian Fe Poland Hungary Czech Rep Austria Switz.Liec Sweden Norway Ireland Finland Denmark Peru Chile Ecuador Venezuela Colombia Oman Qatar Kuwait Taiwan Untd Arab Saudi Arab Philippine New Zealan Pakistan Iran Tunisia Morocco Nigeria Yugoslavia Algeria Israel Kazakhstan Slovenia Slovakia Romania Bulgaria Lebanon Syria Jordan Ukraine Egypt Viet Nam Sri Lanka Bangladesh Croatia Malta Cyprus Ghana Cote Divoi Zimbabwe Iceland Latvia Lithuania Estonia Uzbekistan Belarus TFYR Maced Bosnia Her Uruguay Guatemala Panama Costa Rica Korea D P Cuba Liberia St.Kt-Nev- Bahamas Paraguay Bolivia Honduras El Salvado Neth.Ant.A Jamaica Trinidad T Dominican Myanmar Papua N.Gu Cambodia New Caledn Kiribati Fiji Nepal Lao P.Dem. Libya Iraq Yemen Bahrain Tanzania Kenya Sudan Ethiopia Uganda Malawi Mozambique Madagascar Zambia Mauritius Guinea Togo Mali Burkina Fa Benin Senegal Cameroon Angola Gabon Congo Guadeloupe Fr.Guiana Samoa Oth.Oceani Somalia St.Helena Falkland I St.Pierre Fr Ind O Armenia Afghanista Mongolia Kyrgyzstan Tajikistan Turkmenist Azerbaijan Albania Rep Moldov Georgia Bermuda Guyana Belize Suriname Nicaragua Haiti Barbados Niger Eq.Guinea Greenland Rwanda Burundi Chad Cent.Afr.R GuineaBiss Sierra Leo Gambia Djibouti Seychelles Dem.Rp.Con Mauritania Gibraltar USA Japan Spain Netherland Belgium-Lu UK Germany Italy France,Mon Korea Rep. China India Thailand Indonesia Australia Singapore Malaysia South Afri Portugal Canada Brazil Mexico Argentina Turkey Greece Russian Fe Poland Hungary Czech Rep Austria Switz.Liec Sweden Norway Ireland Finland Denmark Peru Chile Ecuador Venezuela Colombia Oman Qatar Kuwait Taiwan Untd Arab Saudi Arab Philippine New Zealan Pakistan Iran Tunisia Morocco Nigeria Yugoslavia Algeria Israel Kazakhstan Slovenia Slovakia Romania Bulgaria Lebanon Syria Jordan Ukraine Egypt Viet Nam Sri Lanka Bangladesh Croatia Malta Cyprus Ghana Cote Divoi Zimbabwe Iceland Latvia Lithuania Estonia Uzbekistan Belarus TFYR Maced Bosnia Her Uruguay Guatemala Panama Costa Rica Korea D P Cuba Liberia St.Kt-Nev- Bahamas Paraguay Bolivia Honduras El Salvado Neth.Ant.A Jamaica Trinidad T Dominican Myanmar Papua N.Gu Cambodia New Caledn Kiribati Fiji Nepal Lao P.Dem. Libya Iraq Yemen Bahrain Tanzania Kenya Sudan Ethiopia Uganda Malawi Mozambique Madagascar Zambia Mauritius Guinea Togo Mali Burkina Fa Benin Senegal Cameroon Angola Gabon Congo Guadeloupe Fr.Guiana Samoa Oth.Oceani Somalia St.Helena Falkland I St.Pierre Fr Ind O Armenia Afghanista Mongolia Kyrgyzstan Tajikistan Turkmenist Azerbaijan Albania Rep Moldov Georgia Bermuda Guyana Belize Suriname Nicaragua Haiti Barbados Niger Eq.Guinea Greenland Rwanda Burundi Chad Cent.Afr.R GuineaBiss Sierra Leo Gambia Djibouti Seychelles Dem.Rp.Con Mauritania Gibraltar
  • V. Batagelj

Social networks visualization

slide-16
SLIDE 16

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Pathfinder skeleton of World Trade 1999 network

Afghanistan Albania Algeria Angola Argentina Armenia Australia Austria Azerbaijan Bahamas Bahrain Bangladesh Barbados Belarus Belgium-Lux Belize Benin Bermuda Bolivia Bosnia Herzg Brazil Bulgaria Burkina Faso Burundi Cambodia Cameroon Canada Cent.Afr.Rep Chad Chile China Colombia Congo Costa Rica Cote Divoire Croatia Cuba Cyprus Czech Rep Dem.Rp.Congo Denmark Djibouti Dominican Rp Ecuador Egypt El Salvador Eq.Guinea Estonia Ethiopia Falkland Is Fiji Finland Fr Ind O Fr.Guiana France,Monac Gabon Gambia Georgia Germany Ghana Gibraltar Greece Greenland Guadeloupe Guatemala Guinea GuineaBissau Guyana Haiti Honduras Hungary Iceland India Indonesia Iran Iraq Ireland Israel Italy Jamaica Japan Jordan Kazakhstan Kenya Kiribati Korea D P Rp Korea Rep. Kuwait Kyrgyzstan Lao P.Dem.R Latvia Lebanon Liberia Libya Lithuania Madagascar Malawi Malaysia Mali Malta Mauritania Mauritius Mexico Mongolia Morocco Mozambique Myanmar Nepal Neth.Ant.Aru Netherlands New Calednia New Zealand Nicaragua Niger Nigeria Norway Oman Oth.Oceania Pakistan Panama Papua N.Guin Paraguay Peru Philippines Poland Portugal Qatar Rep Moldova Romania Russian Fed Rwanda Samoa Saudi Arabia Senegal SeychellesSierra Leone Singapore Slovakia Slovenia Somalia South Africa Spain Sri Lanka St.Helena St.Kt-Nev-An St.Pierre Mq Sudan Suriname Sweden Switz.Liecht Syria Taiwan Tajikistan Tanzania TFYR Macedna Thailand Togo Trinidad Tbg Tunisia Turkey Turkmenistan Uganda UK Ukraine Untd Arab Em Uruguay USA Uzbekistan Venezuela Viet Nam Yemen Yugoslavia Zambia Zimbabwe

  • V. Batagelj

Social networks visualization

slide-17
SLIDE 17

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Goals

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 get insight not only into the structure and characteristics of a given network, but also into how this structure influences processes going on over the network. We usually need several pictures to present the obtained results. The main tool to deal with large objects is abstraction. In graphs it is usually realized using a hierarchy of partitions. Shrinking or extracting selected classes of the partition we obtain a smaller reduced graph. Small graphs can be presented in their totality and in detail within a single view. In a comprehensive view of large graphs, details become lost – conversely a detailed view can encompass only a part of the graph.

  • V. Batagelj

Social networks visualization

slide-18
SLIDE 18

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Decomposition and parts of network

cut-out reduction local global hierarchy context inter-links

  • V. Batagelj

Social networks visualization

slide-19
SLIDE 19

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Generalized Blockmodeling

A blockmodel consists of structures

  • btained by identifying all units

from the same cluster of the clus- tering C. For an exact definition of a blockmodel we have to be precise also about which blocks produce an arc in the reduced graph and which do not, and of what type [9]. Some types of connections are pre- sented in the figure on the next slide. The reduced graph can be represented by relational matrix, called also image matrix.

  • V. Batagelj

Social networks visualization

slide-20
SLIDE 20

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Block Types

  • V. Batagelj

Social networks visualization

slide-21
SLIDE 21

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Characterizations of Types of Blocks

null nul all 0 ∗ complete com all 1 ∗ regular reg 1-covered rows and columns row-regular rre each row is 1-covered col-regular cre each column is 1 -covered row-dominant rdo ∃ all 1 row ∗ col-dominant cdo ∃ all 1 column ∗ row-functional rfn ∃! one 1 in each row col-functional cfn ∃! one 1 in each column non-null

  • ne

∃ at least one 1

∗ except this may be diagonal

A block is symmetric iff ∀X, Y ∈ Ci × Cj : (XRY ⇔ YRX).

  • V. Batagelj

Social networks visualization

slide-22
SLIDE 22

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Blockmodeling as a clustering problem

The goal of blockmodeling is to reduce a large, potentially inco- herent network to a smaller com- prehensible structure that can be interpreted more readily. Block- modeling, as an empirical proce- dure, is based on the idea that units in a network can be grouped according to the extent to which they are equivalent, according to some meaningful definition of equivalence. In model nodes we can pre- serve the information about the subgraph structure: connectivity, acyclic, empty, . . .

  • V. Batagelj

Social networks visualization

slide-23
SLIDE 23

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Large networks

In larger/denser networks there is often too much information to be presented at once. A possible answer are interactive layouts on the computer screen where the user controls what (s)he wants to see. The computer screen is a different medium, which offers many new possibilities: parallel views (global and local); brushing and linking; zooming and panning; temporary elements (additional information about the selected elements, labels, legends, markers, etc.); highlighted selections; and others. These features can and should be maximally leveraged to support data analytic tasks; or repeating the Shneiderman’s mantra: overview first, zoom and filter, then details

  • n-demand (extended with: relate, history and extract).
  • V. Batagelj

Social networks visualization

slide-24
SLIDE 24

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Sheet of paper

The literature on graph drawing, is dominated by the ’sheet of paper’ paradigm – the solutions and techniques are mainly based on the assumption that the final result is a static picture on a sheet of

  • paper. In this case to present a large data set we need a large ’sheet
  • f paper’ – but this has a limit.

In an interactive dynamic visualization of a graph on the computer screen it needs not to be displayed in its totality. Inspecting a visualization the user can select which parts and elements will be displayed and how.

  • V. Batagelj

Social networks visualization

slide-25
SLIDE 25

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Layouts

The basic steps in graph/network visualization are: graph/network → analyses → layouts → viewer → pictures The networks to be visualized are usually not large – up to some thousans of nodes. On this scheme the development of different tools can be based depending on the kind of users (simple, advanced) and their tasks (exploration, monitoring, analysis, reporting, learning, story telling) they address. In some cases already a standard viewer will be sufficient (for example SVG viewer, X3D viewer, or a special graph layout viewer), in others a complete network analysis system is needed. Layouts are obtained by augmenting the network data with results of analyses and user’s decisions to be used to visualize the network. In Pajek’s input format there are several layout elements from Pajek’s predecessors (see Pajek’s manual, pages 69-73).

  • V. Batagelj

Social networks visualization

slide-26
SLIDE 26

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Styles

As in typesetting text + formatting = formated text so in network visualization network + layout = picture It would be useful to define a common layout format (an extension of GraphML or JSON ?) so that independent viewer modules can be developed and combined with different layout algorithms. To specify layouts we can borrow from the typesetting the notion of style. Some useful ideas can be found in nViZn (”envision”) system [14]. Another interesting option are network visualization solutions based

  • n the library d3.js (Javascript and SVG).
  • V. Batagelj

Social networks visualization

slide-27
SLIDE 27

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Visual complexity

  • V. Batagelj

Social networks visualization

slide-28
SLIDE 28

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Visual complexity

  • V. Batagelj

Social networks visualization

slide-29
SLIDE 29

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Glasses: Rasmol displays

BallStick, SpaceFill, Backbone, Ribbons

  • V. Batagelj

Social networks visualization

slide-30
SLIDE 30

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Glasses, lenses and zooming

Glasses have effect on the entire window, and lenses only on the selected region or elements. Selecting different glasses we obtain different views on the same data – supporting different visualisation aims. 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 nodes only, selecting the type of labels (long/short name, value), displaying only the important nodes and/or links, size of nodes determined by core number or betweenness. In the two pictures produced by James Moody the glasses are the coloring of its nodes by different partitions: age partition (left picture) and race partition (right picture).

  • V. Batagelj

Social networks visualization

slide-31
SLIDE 31

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Part of the big picture (EAT)

The glasses in this case are based on ordering the edges in increasing order of their values and drawing them in this order – stronger edges cover the weaker. The picture emphasizes the strongest substructures; the remaining elements form a background.

  • V. Batagelj

Social networks visualization

slide-32
SLIDE 32

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Lenses: temporary info about the selected vertex

An example of lens is presented on the next slide – contributions of companies to different presidential candidates from Follow the Oil

  • Money. For a selected node the information about it is displayed.

Another example of lens would be to temporarily enhance the display

  • f neighbors of the selected node or to display their labels. The

”shaking” option used in Pajek to visually identify all vertices from selected cluster is also a kind of lense. Additional enhancement of a presentation can be achieved by the use

  • f support elements such as labels, grids, legends, and various forms
  • f help facilities.

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

  • V. Batagelj

Social networks visualization

slide-33
SLIDE 33

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Lenses: temporary info about the selected vertex

  • V. Batagelj

Social networks visualization

slide-34
SLIDE 34

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Zoom, glasses, lenses, navigation: Google Maps

A nice example of combination of these techniques is the Google Maps

  • service. It combines zooming, glasses (Map, Satellite, Terrain), navigation

(left, right, up, down) and lenses (info about points). The maps at different zoom levels provide information at different level of detail and in different form.

  • V. Batagelj

Social networks visualization

slide-35
SLIDE 35

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Zoom, glasses, lenses, navigation: Grokker

  • V. Batagelj

Social networks visualization

slide-36
SLIDE 36

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Maps

For a given hierarchical clustering of its nodes a similar approach to Grokker could be used also for inspection of large graphs/networks. To produce higher level ’maps’ different methods can be used: k-core representation, density contours, generalized blockmodeling, clustering, skeletons preserving only important nodes and links, etc. In visualizing the ’maps’ new graphical elements (many of them still to be invented) can be used preserving/indicating the information about the structure at lower level.

  • V. Batagelj

Social networks visualization

slide-37
SLIDE 37

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

k-core structure of the .fr domain

Alessandro Vespignan et. al.

  • V. Batagelj

Social networks visualization

slide-38
SLIDE 38

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

Problems

1 Network maps:

area photo → map; network → ? ;

2 Streams in acyclic networks:

AcyNet → Map → layout of AcyNet based on Map;

3 Description of layouts:

data → NA → layout → viewer(s) → picture(s).

  • V. Batagelj

Social networks visualization

slide-39
SLIDE 39

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

References I

Abello, J., van Ham, F.: Matrix Zoom – A Visual Interface to Semi-external Graphs. IEEE Symposium on Information Visualization 2004, October 10-12, Austin, Texas, USA, 183-190. Abello, J., van Ham, F., Neeraj Krishnan: ASK-GraphView : A Large Scale Graph Visualization System. IEEE Transactions on Visualization and Computer Graphics, 12 (2006) 5. Batagelj, V.: Complex Networks, Visualization of. In R.A. Meyers, ed., Encyclopedia of Complexity and Systems Science, Springer 2009: 1253-1268. Batagelj, V.: Some visualization challenges from SNA. Graph Drawing 2005 workshop slides. Limerick, 2005. Batagelj, V., Doreian, P., Ferligoj, A., Kejˇ zar, N.: Understanding Large Temporal Networks and Spatial Networks. Wiley, 2014.

  • V. Batagelj

Social networks visualization

slide-40
SLIDE 40

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

References II

Batagelj, V., Mrvar, A., Zaverˇ snik, M.: Partitioning approach to visualization of large graphs, In Kratochv´ ıl, J. (ed), Lecture notes in computer science, 1731. Springer, Berlin, 1999, 90–97. Vladimir Batagelj, Andrej Mrvar: Pajek manual. Brandes, U., Pich, C.: Eigensolver Methods for Progressive Multidimensional Scaling of Large Data. In: Proceedings of Graph Drawing ‘06 (LNCS 4372), 2007, pp. 42-53. Doreian, P., Batagelj, V., Ferligoj, A.: Generalized Blockmodeling. Cambridge University Press, 2005. Henry, N., Fekete, J-D.: MatrixExplorer: a Dual-Representation System to Explore Social Networks. IEEE Transactions on Visualization and Computer Graphics, 12 (2006) 5.

  • V. Batagelj

Social networks visualization

slide-41
SLIDE 41

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

References III

Kobourov, S.G.: Force-Directed Drawing Algorithms. In R. Tamassia: Handbook of Graph Drawing and Visualization. CRC, 2013, pp. 383-408. Mueller, C.: Matrix visualizations. http://www.osl.iu.edu/~chemuell/data/ordering/sparse.html Mueller, C., Martin, B., Lumsdaine, A.: A Comparison of Vertex Ordering Algorithms for Large Graph Visualization. APVIS 2007. SPSS nViZn: http://www.spss.com/research/wilkinson/nViZn/nvizn.html Yifan Hu: Gallery of Large Graphs. http://www.research.att.com/~yifanhu/GALLERY/GRAPHS/index1.html De Nooy, W., Mrvar, A., Batagelj, V.: Exploratory Social Network Analysis with Pajek; Revised and Expanded Second Edition. Cambridge University Press, September 2011.

  • V. Batagelj

Social networks visualization

slide-42
SLIDE 42

Social networks visualization

  • V. Batagelj

Networks Approaches Statistics Drawing of large graphs Drawing of dense graphs Network decomposition Layouts Problems References

References IV

Waltman, L., van Eck, N.J., Noyons, E.C.M.: A unified approach to mapping and clustering of bibliometric networks. Journal of Informetrics, 4(2019)4, pp. 629–635. Wilkinson, L.: The Grammar of Graphics. Springer-Verlag, New York, 2005.

  • V. Batagelj

Social networks visualization