Upcoming SlideShare
×

# Network Visualization guest lecture at #DataVizQMSS at @Columbia / #SNA at PUC Chile

14,325 views

Published on

- First version was a guest lecture about Network Visualization in the class "Data Visualization" taught by Dr. Sharon Hsiao in the QMSS program at Columbia University http://www.columbia.edu/~ih2240/dataviz/index.htm
- This updated version was delivered in our class on SNA at PUC Chile in the MPGI master program.

40 Likes
Statistics
Notes
• Full Name
Comment goes here.

Are you sure you want to Yes No
• Be the first to comment

Views
Total views
14,325
On SlideShare
0
From Embeds
0
Number of Embeds
938
Actions
Shares
0
267
0
Likes
40
Embeds 0
No embeds

No notes for slide
• 9*8 = 72, 72 / 2 = 36
• 94*93 = 8742Density: Network density is the proportion of edges in a network relative to the total number of possible edges.Diameter: The diameter of a network is the length (#edges) of the longest path between two nodesClustering Coefficient: : A measure of the likelihood that two associates of a node are associates themselves. A higher clustering coefficient indicates a greater cliquishness.Path Length: The distances between pairs of nodes in the network. Average path-length: is the average of these distances between all pairs of nodes.
• 68*67=4556
• Visualizations that use Network Visualization
• ### Network Visualization guest lecture at #DataVizQMSS at @Columbia / #SNA at PUC Chile

1. Network  Visualiza0on   +  Gephi  Tutorial   Denis  Parra,  Ph.D.   Assistant    Professor   Pon0ﬁcal  Catholic  University  of  Chile   MPGI  ~  Master  Program   Monday,  October  5th,  2015
2. Expecta0ons   •  What  will  you  learn  at  the  end  of  this  class?   1.  Basic  concepts  of  Networks,  Graphs,  and  Social   Network  Analysis  (SNA)   2.  Systems/Applica0ons  that  make  use  of  network   visualiza0ons   3.  Recent  Research  on  Network  Visualiza0on   4.  How  to  use  a  Network  Visualiza0on  and  Analysis     tool  (Gephi)  ~  in  class  tutorial   5.  Bonus  :  Where  do  I  ﬁnd  data  sets  to  do  more   cool  visualiza0ons?   10/5/15   @denisparra  |    2
3. 1.  Basic  Concepts  and   Deﬁni0ons
4. We  live  in  a  connected  world   •  …  and  we  need  visualiza0on  models  to  represent   networks  such  as:   –  Online  Social  networks:  Facebook,  Twiaer  ~  people   connected  online   –  Informa3on  networks:  WWW  ~  web  pages  connected   through  hyperlinks   –  Computer  networks:  The  internet  ~  computers  and   routers  connected  through  wired/wireless  connec0ons     •  What  is  a  network?  (Easley  and  Kleinberg,  2011)   “a  network  is  any  collec0on  of  objects  in  which   some  pairs  of  these  objects  are  connected  by   links”.   10/5/15   @denisparra  |    4
5. A  bit  of  history:  Graph  models   •  Around  1735,  the  mathema0cian  Venn  Euler   set  the  founda0on  for  graph  theory  by   crea0ng  a  model  to  represent  the  problem  of   the  “7  bridges  of  Königsberg”       Source:  hap://en.wikipedia.org/wiki/Seven_Bridges_of_K%C3%B6nigsberg  and  “Linked”   by  A-­‐L.  Barabasi   10/5/15   @denisparra  |    5
6. A  formal  deﬁni0on  of  Graph   •  Based  on  (Easley  and  Kleinberg,  2011):  A   graph  is  a  way  of  specifying  rela.onships   among  a  collec.on  of  items.  A  graph   consists  of  a  set  of  objects,  called  nodes,   with  certain  pairs  of  these  objects   connected  by  links  called  edges.   10/5/15   @denisparra  |    6
7. Graphs  as  Models  of  Networks   •  Based  on  (Easley  and  Kleinberg,  2011):   “Graphs  are  useful  because  they  serve  as   mathema0cal  models  of  network  structures.”   •  But  keep  in  mind:  Graphs  are  only  one  way  to   represent  networks  (though  the  most   popular)   {  arc  ,  link  ,  edge  }   {  node  ,  vertex  }   Source:  L.  Adamic  SNA  class  @coursera   10/5/15   @denisparra  |    7
8. The  Historic  Development  of  Network   Visualiza0on   •  The  following  slides  are  based  on  the  work  of   Pfeﬀer  and  Freeman  (2015)   Pfeﬀer,   Juergen   &   Freeman,   Lin.   Methods   of   Social   Network  VisualizaAon.  Encyclopedia  of  Complexity  and   Systems  Science,  2nd  EdiAon,  Springer  Reference.   •  They  categorize  this  historic  development  on   three  categories:   1.  Nodes,  Links,  Shape,  Size   2.  Substance-­‐Based  Layout   3.  Two-­‐Mode  Networks   10/5/15   @denisparra  |    8
9. Overall  View  of  the  Visualiza0ons     Reference:     Pfeﬀer,  Juergen  &   Freeman,  Lin   (forthcoming).  Methods   of  Social  Network   Visualiza0on.   Encyclopedia  of   Complexity  and  Systems   Science,  2nd  Edi0on,   Springer  Reference.     hap://www.pfeﬀer.at/ data/visposter/   @denisparra  |    9
10. 1.  Nodes,  Links,  Shape,  Size  (1/2)     @denisparra  |    10
11. 1.  Nodes,  Links,  Shape,  Size  (2/2)     @denisparra  |    11
12. 2.  Substance-­‐Based  Layout  (1/2)     @denisparra  |    12
13. 2.  Substance-­‐Based  Layout  (2/2)     @denisparra  |    13
14. 3.  Two-­‐mode  Networks  (1/2)     @denisparra  |    14
15. 3.  Two-­‐mode  Networks  (2/2)     @denisparra  |    15
16. The  tennis  players’  social  network   Sharonpova                          Sharapova                                            Serena                                          Li  Na          Rafa                                                    Djoker                                                Soderling                        Prof.  Parra   Roger   10/5/15   @denisparra  |    16
17. Some  Types  of  Networks   Undirected   (Facebook   friendships)   Directed   (TwiAer   following)   mul3mode   (Amazon  user-­‐ product)   Weighted   (Facebook    likes)   and   more ….   9   3   •  Hereinauer,  I  will  refer  indis0nc0vely  to   graphs  and  networks.  Here  some  types:     10/5/15   @denisparra  |    17
18. Analyzing  a  network:  SNA   •  How  do  we  analyze  a  network?   •  How  do  we  compare  diﬀerent  networks?   •  This  class  is  about  network  visualiza0ons,  but   some  founda0onal  concepts  of  SNA  need  to   be  understood  before.   •  Let’s  see  ways  to  describe  the  network  at   local  and  at  global  level   Source:  hap://moviegalaxies.com   10/5/15   @denisparra  |    18
19. Measures  in  SNA   Node-­‐level  metrics   •  Centrality   –  (In/Out)  Degree   –  Betweenness   –  Closeness   –  Eigenvector   •  Clustering  coeﬃcient   Graph-­‐level  metrics   •  Size   •  Diameter  (longest   path)   •  Average  path  length   •  Average  [node  metric]   10/5/15   @denisparra  |    19   •  These  are  only  a  few  representa0ve  measures   •  For  further  understanding  of  these  measures:  See  the   presenta0on  of  Giorgos  Chelo0s  in  slideshare,  from  slide  8     hap://www.slideshare.net/gchelio0s/social-­‐network-­‐analysis-­‐3273045
20. Interpreta0on  of  measures   10/5/15   @denisparra  |    20   Source:  hap://www.slideshare.net/gchelio0s/social-­‐network-­‐analysis-­‐3273045  slide  24   Interpreta3on  in  Social  Networks   Degree   How  many  people  can  this  person   reach  directly?   Betweenness   How  likely  is  this  person  to  be  the   most  direct  route  between  two   people  in  the  network?
21. Interpreta0on  of  measures   10/5/15   @denisparra  |    21   Source:  hap://www.slideshare.net/gchelio0s/social-­‐network-­‐analysis-­‐3273045  slide  24   Interpreta3on  in  Social  Networks   Closeness   How  fast  can  this  person  reach   everyone  in  the  network?   Eigenvector   How  well  is  this  person  connected   to  other  well-­‐connected  people?
22. Two  more  concepts…   •  Total  possible  number  of  edges  in  a  network   #edges  =  n  *  (n  -­‐1  )  /2  (undirected  network)   #edges  =  n  *  (n  -­‐1  )  (directed  network)   •  (Shortest)  Path:  the  shortest  sequence  of   edges  to  be  followed  to  reach  a  node  B  from  a   node  A  in  a  network.   Which  is  the  length  of  the  shortest  path     between  Rafa  Nadal  and  Sharonpova?   10/5/15   @denisparra  |    22
23. Prac0ce  the  learned  concepts…   •  Prac0ce  the  learned  concepts  comparing   these  2  movie  networks  (characters’   interac0ons)  :   Traﬃc  (2000)  Forrest  Gump  (1994)   Source:  hap://moviegalaxies.com   10/5/15   @denisparra  |    23
24. Forrest  Gump  (1994)   Network  metrics:   •  Size:  94/271   •  Density:  0.06   •  Diameter:  4   •  Clustering  coeﬃcient:   0.8   •  Avg.  Path  Length:  1.99     Node  metrics:   Forrest   •  Degree:  89   •  Betweetnness:  3453.8   Abbie  Hoﬀman   •  Degree:  6   •  Betweenness:  0   hap://moviegalaxies.com/movies/316-­‐Forrest-­‐Gump   Abbie  Hoﬀman   10/5/15   @denisparra  |    24
25. Traﬃc  (2000)   Network  metrics:   •  Size:  68   •  Density:  0.04   •  Diameter:  7   •  Clustering  coeﬃcient:   0.55   •  Avg.  Path  Length:  3.54     Node  metrics:   Robert   •  Degree:  24   •  Betweetnness:  1437.7   Francisco   •  Degree:  5   •  Betweenness:  1031   hap://moviegalaxies.com/movies/837-­‐Traﬃc   *  Francisco:  is  a  bridge  (structural  holes)   10/5/15   @denisparra  |    25
26. Network  Components   •  (from  G.  Chelio0s)  “many  large  groups   and  online  communi0es  have  a  core  of   densely  connected  users  …  and  a  much   larger  periphery”   •  Source:     hap://www.slideshare.net/gchelio0s/social-­‐network-­‐analysis-­‐3273045,   page  34   •  (from  L.  Adamic)  “if the largest component encompasses a significant fraction of the graph, it is called the giant component” •  Source:     haps://class.coursera.org/sna-­‐2012-­‐001/class/index  ,  week  1  slides   10/5/15   @denisparra  |    26
27. Remarks  and  Further  topics  in  SNA   •  With  the  concepts  already  described,  we  will  aaempt  to   visualize  and  analyze  two  networks  in  the  NodeXL  &  Gephi   tutorial.   •  Not  covered  in  this  class,  but  worth  men0oning  other  SNA   topics:   –  Network  growth/forma0on:  Erdős–Rényi,  Waas-­‐Strogatz,   Barabassi-­‐Albert  (preferen0al  aaachment)   –  Community  Structure:  Girvan-­‐Newman,  Clauset-­‐Moore-­‐ Newman  (max-­‐modularity),  aﬃnity  propaga0on,  etc.   –  Processes  in  networks:  Diﬀusion,  epidemics,  innova0on,   etc.   –  Network  mo0fs:    small  subgraphs  that  are  over-­‐represented   in  the  network  10/5/15   @denisparra  |    27
28. 2.  Applica0ons
29. Examples  of  Applica0ons   •  These  are  a  few  examples  of  applica0ons  that   make  use  of  Network  Visualiza0ons:   –  Truthy   –  Moviegalaxies   –  Poderopedia   –  TwiaerScope   –  LinkedIn  Maps   •   These  ARE  NOT  tools  for  generic  Visualiza0on   and  Analysis  (we’ll  see  those  in  the  tutorial   sec0on)   10/5/15   @denisparra  |    29
30. Truthy   •  Informa0on  Diﬀusion  research  at  Indiana  U.   •  hap://truthy.indiana.edu   10/5/15   @denisparra  |    30
31. MovieGalaxies   •  Visualize  an  discuss  the  characters  of  movies   as  networks   •  hap://moviegalaxies.com   10/5/15   @denisparra  |    31
32. Poderopedia   •  Who  is  who  in  business  and  poli0cs  in  Chile?   •  Knight  Founda0on:  Top  10  digital  tools  for   journalists  (Feb  4,  2013)   hap://www.knigh•ounda0on.org/blogs/knightblog/2013/2/4/new-­‐ digital-­‐tools-­‐journalists-­‐10-­‐learn/   10/5/15   @denisparra  |    32
33. TwiaerScope   •  A  visual  monitor  of  tweets  in  real  0me.  This  is   an  enhanced  graph  model.   •  hap://0bes0.research.aa.com/twiaerscope/   10/5/15   @denisparra  |    33
35. 3.  Recent  Research
36. Recent  Research  (~by  Feb  2013)   •  Can  we  go  Beyond  the  Graph?   •  ManyNets   •  HivePlots   •  Orion   •  GraphPrism   •  Mo0f  Simpliﬁca0ons   •  GeoSpa0al  Network  Visualiza0on   10/5/15   @denisparra  |    36
37. Social  Network  Visualiza0on:  Can  we   go  Beyond  the  Graph?  (2006)   •  Authors  support  that  social  network   visualiza0on  for  end  users  should  go  beyond   the  graph-­‐only  paradigm   •  hap://web.media.mit.edu/~fviegas/papers/viegas-­‐cscw04.pdf   10/5/15   @denisparra  |    37
38. ManyNets  (2010)   •  Analyze  and  compare  mul0ple  networks   •  hap://www.cs.umd.edu/hcil/manynets/   10/5/15   @denisparra  |    38
39. Hive  Plots  (2011)   •  Hive  plots—ra0onal  approach  to  visualizing   networks   hap://www.hiveplot.net/   10/5/15   @denisparra  |    39
40. Orion  (2011)   •  Diﬀerent  visualiza0ons  to  present  network   data   •  hap://vis.stanford.edu/papers/orion   a)  Sorted  matrix   b)  Node-­‐link  diagram   c)  Plot  of  betweenness   for  two  networks   10/5/15   @denisparra  |    40
41. GeoSpa0al  Network  Visualiza0on   (2011)   •  Interac0ve  Explora0on  of  Geospa0al  Network   Visualiza0on   •  hap://0llnagel.com/2011/10/interac0ve-­‐ explora0on-­‐of-­‐geospa0al-­‐network-­‐ visualiza0on/   10/5/15   @denisparra  |    41
42. GraphPrism  (2012)   •  GraphPrism:  Compact  Visualiza0on  of   Network  Structure,  inspired  in  B-­‐Matrices   •  hap://vis.stanford.edu/papers/graphprism   10/5/15   @denisparra  |    42
43. Mo0f  Simpliﬁca0on  (2012)   •  Use  of  fans  and  parallel  glyphs  to  improve  readibility   •  hap://hcil2.cs.umd.edu/trs/2012-­‐11/2012-­‐11.pdf     10/5/15   @denisparra  |    43
44. MuxViz:  Mul0layer  Networks  (2014)   •  Mul0layer  analysis  and  visualiza0on  of  networks   •  hap://muxviz.net/index.php   10/5/15   @denisparra  |    44
45. 4.  Using  a  Network   Visualiza0on  Tool   (NodeXL  &  Gephi  in  a  nutshell)
46. Network  Analysis  and  Visualiza0on   Tools   •  NodeXL   •  Gephi   •  Pajek   •  ORA  (CMU)   •  igraph  (C++,  R)   •  UCINet   •  NetworkX   •  Tulip   •  Visone   •  Larger  list:  hap://www.gmw.rug.nl/~huisman/sna/ souware.html   10/5/15   @denisparra  |    46
47. How  do  I  format  my  network  data?   •  Depends  on  your  informa0on  needs.  What  do  you   want  to  describe?   – GDF  hap://guess.wikispot.org/The_GUESS_.gdf_format   – GEXF  hap://gexf.net/format/   – GraphML  hap://graphml.graphdrawing.org   – Pajek  Net  format   hap://vlado.fmf.uni-­‐lj.si/pub/networks/pajek/doc/pajekman.pdf   – CSV  haps://gephi.org/users/supported-­‐graph-­‐formats/csv-­‐format/   •  For  a  summary  and  examples,  check   haps://gephi.org/users/supported-­‐graph-­‐formats/   10/5/15   @denisparra  |    47
48. How  do  I  format  my  Data?   10/5/15   @denisparra  |    48
49. For  the  rest  of  my  classes   •  Will  use  this  as  reference  for  iGraph  analysis
50. Gephi  tutorial   hap://web.ing.puc.cl/~dparra/classes/sna-­‐ INP3460-­‐2015-­‐2/NetworkViz-­‐tutorial-­‐ instruc0ons.pdf     10/5/15   @denisparra  |    50
51. Dynamic  Network  Visualiza0on   •  Data:  hap://web.ing.puc.cl/~dparra/classes/ sna-­‐INP3460-­‐2015-­‐2/data/ bipar0te_nodes_CSCW2013_URL_0me.gdf   •  Hot-­‐to-­‐visualize:   haps://kawinproject.wordpress.com/ 2013/03/10/dynamic-­‐data-­‐longitudinal-­‐ network-­‐in-­‐gephi/
52. Final  Remarks   •  In  this  class  you  learnt:   –  Basic  concepts  of  networks,  graphs,  and  SNA   –  Existent  applica0ons  that  make  use  of  network   visualiza0ons   –  Research  related  to  network  visualiza0on   –  How  to  use  a  network  visualiza0on  and  analysis  tool   •  My  ﬁnal  message:     –  Graph  model  is  great,  but  try  to  move  beyond  the   graph-­‐only  visualiza0on.     –  Think  of  ways  to  create  visualiza0ons  that  help  to   make  sense  of  the  diﬀerent  proper0es  inherent  to  the   network  and  to  its  elements  (nodes  and  links).  R  and   Javascript  give  you  enough  power  to  implement.   10/5/15   @denisparra  |    52
53. Thanks!   •  Ques0ons?   •  denisparra@gmail.com  or  @denisparra   •  Check  my  academic  web  page   hap://web.ing.puc.cl/~dparra/   •  and  my  research  blog     hap://kawinproject.wordpress.com
54. 5.  Bonus  Slides
55. Where  do  I  ﬁnd  cool  NetVis?   •  hap://www.visualcomplexity.com/vc/   Where  do  I  ﬁnd  network  datasets?   •  Jure  Leskovec  page  hap://snap.stanford.edu/data/   •  Mark  Newman’s  page  hap://www-­‐personal.umich.edu/~mejn/netdata/   •  Gephi  wiki  datasets  hap://wiki.gephi.org/index.php/Datasets   •  From  CMU’s  Graphlab  hap://graphlab.org/downloads/datasets/   10/5/15   @denisparra  |    55
56. Recommended  books   •  Linked  by  Albert  L.  Barabasi   •  Networks,  Crowds,  and  Markets  by  D.  Easley   and  J.  Kleinberg  (pre-­‐print  available  free   online)   10/5/15   @denisparra  |    56
57. Recommended  Online  Tutorials   •  Gephi:   – At  ICWSM  ‘11   hap://www.slideshare.net/Cloud/sp1-­‐ exploratory-­‐network-­‐analysis-­‐with-­‐gephi   – Gephi  online  tutorial  hap://blog.ouseful.info/ 2012/11/09/drug-­‐deal-­‐network-­‐analysis-­‐with-­‐ gephi-­‐tutorial/#   – Lada  Adamic  2012  SNA  class:     hap://www.youtube.com/watch? v=JgDYV5ArXgw&list=PL828B49781EAA17ED   10/5/15   @denisparra  |    57
58. •  Do  you  R?   – Temporal  networks  with  igraph  and  R  (with  20   lines  of  code!)   hap://markov.uc3m.es/2012/11/temporal-­‐ networks-­‐with-­‐igraph-­‐and-­‐r-­‐with-­‐20-­‐lines-­‐of-­‐ code/   10/5/15   @denisparra  |    58
59. LineSets  (InfoVis  2011)   •  Alper  et  al.  (UCSB  and  Microsou  research)   •  Extend  a  concept  from  subway  maps  to  sets  of   items   10/5/15   @denisparra  |    59
60. Denis  Parra’s  Research   •  Using  networks  vis.  in  recommenda0on   approaches:  “Visualizing  Recommenda3ons   to  Support  Explora3on,  Transparency  and   Controllability”  by  Verbert,  Parra,  Brusilovsky   and  Duval,  IUI  Conference  (2013).
61. Denis  Parra’s  Research   •  Plo‡ng  edges’  weight  distribu0ons  of  several   networks  to  compare  community  explain   algorithms  performance   10/5/15   @denisparra  |    61
62. Denis  Parra’s  Research   •  Twiaer  in  academic  events:  A  study  of   temporal  usage,  communica0on,  sen0mental   and  topical  paaerns  in  16  Computer  Science   conferences  ( hap://dx.doi.org/10.1016/j.comcom. 2015.07.001  )