Guillaume Er

1
analysis of a real online social network using
semantic web frameworks

• Guillaume Erétéo, Michel Buffa, Fabien Gandon,
  Olivier Corby

2
computer-mediated networks as social networks
Wellman, 2001
3
social media landscape

• social web amplifies social network effects

4
overwhelming flow of social data
consulting
notifying
animating
monitoring
5
social network analysis

• proposes graph algorithms to characterize the
  structure of a social network, strategic
  positions, and networking activities

6
social network analysis

• global metrics and structure

density and diameter cohesion of the network
community detection distribution of actors and
activities
7
social network analysis

• strategic positions and actors

degree centrality local attention
8
social network analysis

• strategic positions and actors

betweenness centrality reveal broker "A place for
good ideas" Burt, 2004
9
semantic social networks
http//sioc-project.org/node/158
10
Fabien
Mylène
Gérard
colleague
father
sister
(guillaume)5
d
colleague
mother
Michel
Yvonne
11
Fabien
Mylène
Gérard
colleague
father
sister
d
(guillaume)3
ltfamilygt
colleague
mother
Michel
Yvonne
12
but

• SPARQL is not expressive enough to meet SNA
  requirements for global metric querying of social
  networks (density, betweenness centrality, etc.).

San Martin Gutierrez 2009
13
classic SNA on semantic web

• rich graph representations reduced to simple
• untyped graphs

Paolillo Wright, 2006
foafknows
foafinterest
14
semantic SNA stack

• exploit the semantic of social networks

15
SPARQL extensions

• CORESE semantic search engine implementing
  semantic web languagesusing graph-based
  representations

16
grouping results

• number of followers of a twitter user
• select ?y count(?x) as ?indegree where
• ?x twitterfollow ?y
• group by ?y

17
path extraction

• people knowing, knowing, (...) colleagues of
  someone
• ?x sa (foafknows/relworksWith)path ?y
• filter(pathLength(path) lt 4)
• Regular expression operators are / (sequence)
  (or) (0 or more) ? (optional) !
  (not)
• Path characteristics i to allow inverse
  properties, s to retrieve only one shortest path,
  sa to retrieve all shortest paths.

18
full example

• closeness centrality through knows and worksWith
• select distinct ?y ?to
• pathLength(path) as ?length
• (1/sum(?length)) as ?centrality
• where
• ?y s (foafknows/relworksWith)path ?to
• group by ?y

19
Qualified component
Qualified degree
Qualified in-degree
Qualified diameter
Number of geodesics between from and to
Number of geodesics between from and to going
through b
Closenness Centrality
Betweenness Centrality
20
SemSNA an ontology of SNA

• http//ns.inria.fr/semsna/2009/06/21/voc

21
add to the RDF graph

• saving the computed degrees for incremental
  calculations
• CONSTRUCT
• ?y semsnahasSNAConcept _b0
• _b0 rdftype semsnaDegree
• _b0 semsnahasValue ?degree
• _b0 semsnaisDefinedForProperty relfamily
• SELECT ?y count(?x) as ?degree where
• ?x relfamily ?y
• UNION
• ?y relfamily ?x
• group by ?y

22
4
Gérard
Mylène
hasValue
hasCentralityDistance
2
Degree
sister
father
isDefinedForProperty
colleague
hasSNAConcept
Yvonne
mother
Guillaume
supervisor
supervisor
supervisor
Michel
Fabien
colleague
colleague
colleague
colleague
Ivan
Philippe
Peter
23
Ipernity
24
using real data

• extracting a real dataset from a relational
  database
• construct ?person1 relfriendOf ?person2
• select sql(ltservergt, ltdrivergt, ltusergt, ltpwdgt,
  select user1_id, user2_id from
  relations where rel 1 ') as
  (?person1 , ?person2 )
• where

25
importing data with SemSNI

• http//ns.inria.fr/semsni/

26
using real data

• ipernity.com dataset extracted in RDF61 937
  actors 494 510 relationships
• 18 771 family links between 8 047 actors
• 136 311 friend links implicating 17 441 actors
• 339 428 favorite links for 61 425 actors
• 2 874 170 comments from 7 627 actors
• 795 949 messages exchanged by 22 500 actors

27
performances limits
time
projections
Knows 0.71 s 494 510
Favorite 0.64 s 339 428
Friend 0.31 s 136 311
Family 0.03 s 18 771
Message 1.98 s 795 949
Comment 9.67 s 2 874 170
Knows 20.59 s 989 020
Favorite 18.73 s 678 856
Friend 1.31 s 272 622
Family 0.42 s 37 542
Message 16.03 s 1 591 898
Comment 28.98 s 5 748 340
Shortest paths used to calculate Knows Path length lt 2 14m 50.69s  Path length lt 2 2h 56m 34.13s Path length lt 2 7h 19m 15.18s  100 000 1 000 000 2 000 000
Shortest paths used to calculate Favorite Path length lt 2 5h 33m 18.43s 2 000 000
Shortest paths used to calculate Friend Path length lt 2 1m 12.18 s  Path length lt 2 2m 7.98 s 1 000 000 2 000 000
Shortest paths used to calculate Family Path length lt 2 27.23 s Path length lt 2 2m 9.73 s Path length lt 3 1m 10.71 s Path length lt 4 1m 9.06 s 1 000 000 3 681 626 1 000 000 1 000 000
28
some interpretations

• validated with managers of ipernity.com
• friendOf, favorite, message, commentsmall
  diameter, high density
• family as expected large diameter, low density
• favorite highly centralized around Ipernity
  animator.
• friendOf, family, message, comment power law of
  degrees and betweenness centralities, different
  strategic actors
• knows analyze all relations using subsumption

29
some interpretations

• existence of a largest component in all sub
  networks
• "the effectiveness of the social network at doing
  its job" Newman 2003

30
conclusion

• directed typed graph structure of RDF/S well
  suited to represent social knowledge socially
  produced metadata spanning both internet and
  intranet networks.
• definition of SNA operators in SPARQL (using
  extensions and OWL Lite entailment) enable to
  exploit the semantic structure of social data.
• SemSNAorganize and structure social data.

31
perspectives

• semantic based community detection algorithm
• SemSNA Ontology
• extract complex SNA features reusing past results
• support iterative or parallel approaches in the
  computations
• a semantic SNA to foster a semantic intranet of
  people
• structure overwhelming flows of corporate social
  data
• foster and strengthen social interactions
• efficient access to the social capital Krebs,
  2008 built through online collaboration

http//twitter.com/isicil
32
slideshare.net/ereteog
twitter.com/ereteog
holdsAccount
holdsAccount
name
Guillaume Erétéo
mentorOf
organization
answers
manage
contribute
mentorOf
contribute