A Social Network Intelligence Tool for visual Analysis of...
-
Upload
hoangthuan -
Category
Documents
-
view
219 -
download
0
Transcript of A Social Network Intelligence Tool for visual Analysis of...
© Net Business Center / 1©2007 IKM,SYSEDV, TU Berlin
A Social Network Intelligence ToolA Social Network Intelligence Toolfor visual Analysis offor visual Analysis of
electronic Communication Networks electronic Communication Networks
27.03.2007 (GOR), M. Trier, IKM Research, SYSEDV, TU Berlin
(Kontakt: [email protected])
„Coming together is a beginning; keeping together is progress; working together is success“
Henry Ford
© Net Business Center / 2©2007 IKM, SYSEDV, TU Berlin, / 2
AGENDA
1. Motivation and ObjectiveCommunication and online networks
2. Research ApproachesModeling networks as CommunigraphsApproach of Dynamic Network AnalysisIT-Support: Commetrix Project
3. Applications and BenefitsAlgorithms and Methods applied to a newsgroup the corporate e-mail network of Enron
© Net Business Center / 3©2007 IKM, SYSEDV, TU Berlin, / 3
Content vs. Communication
Content Communication
Social Content
PersistentConversation
CommunityBuilding/
Networking
User generatedcontent
CollaborativeContent Development
(Wiki)
GroupsSocialization
Emergent Structures
Systems
Networks
Collective Intelligence
Micro Publishing(Blogs)
RSS-Feeds
(cf. Trier, 2007, Content und Kommunikation, In: Herzog (Ed.): Medienproduktion, in Press)
© Net Business Center / 4©2007 IKM, SYSEDV, TU Berlin, / 4
The changing world of online communication: Communication and Networks
Electronic communication:
novel combinatorial forms of content and communication.
micro-publishing, social contents, or collaborative content development.
Larger groups of people contribute user-generated content.
(read-and-write-web, bring-your-own-content)
increased perception of the link between contents and authors.
users connect, form allies and as a group support each other.
giving advice, directing the attention, or by collectively filtering
How can virtual social architectures be captured, modeled and understood?
*cf. Source: Trier (2007): Content und Kommunikation. In: Herzog (2007): Medienproduktion. In press.
*
*
© Net Business Center / 5©2007 IKM, SYSEDV, TU Berlin, / 5
Application of Social Network Analysis
Definition Social Network Analysis (SNA):Framework for the analysis of structured social relationships [WaFa94]; Authority relationships, informal communication, information exchange, affection
Current Measures:Size, relationship strength, roles (broker, hub, isolate, transmitter), activity, prominence, symmetry, centrality,…
Many Fields of Application:Interorganizational collaboration/information and knowledge flow – between core processes, practice areas, hierarchical levels, organizations, teams, departments, projects, partners, expertsTrace important networkers and multiplicatorsMap planned and living processes or structuresNetwork integration after structural change (e.g. merger)Information transfer bottlenecksMapping online communication within social architectures
*cf. Source: Trier (2007): Content und Kommunikation. In: Herzog/Sieck (2007): Medienproduktion, in press.
© Net Business Center / 6©2007 IKM, SYSEDV, TU Berlin, / 6
Shortcomings by current structural outcomes (static SNA)
Questions:Did the actor react to external events?Has the actor a long steady or a short but quick growth in his degree?Is the network stable or eventually already decaying again? Is a cluster stable or just an additive artefact?Which clusters do form and decay?Which are the most important nodes?How can I elicit the core structures? How are the dynamics of adopting a new topic?Where is the locus of novel ideas?How similar is the discussion at different locations in the network?How embedded are certain ego-networks within the overall network?Change can average out between two states of a network and thus can sometimes be poorly captured with statistical comparison
A
B
sending
receiving
© Net Business Center / 7©2007 IKM, SYSEDV, TU Berlin, / 7
Methods: Towards Dynamic Network Visualization and Analysis
labellabellabel
label
label
NodeActors
NodeattributesNames, Function, Locations, Evaluations
EdgeMessages, Contacts, Helprequests, Shared Objects
EdgeattributesEvent/Timestamp, Reference, Types, Content, Topics
Communigraph:Size, color, rings, brightness, position, label
Time window
Time interval
a) cumulative b) time-window/flipchart-book /decay
Model Elements
© Net Business Center / 8©2007 IKM, SYSEDV, TU Berlin, / 8
IT-Support: Commetrix Software Tool
Grafische Modelle und
AuswertungenE-Mail, Discussion, IM, Collab. Tool, VoIP, PMS
Mining & Visualization
Evolution of Virtual Groups
Topic Analysis,Expertise Search &
Expertise Maps
Elicitingcorestructuresand people incomplexnetworks www.commetrix.de
© Net Business Center / 9©2007 IKM, SYSEDV, TU Berlin, / 9
1) Project Initiation
2) Selecting relevant network – defining scope
3) capture and refine data / (multiple sources)
4) Visualization & Analysis
structure
contents
static dynamic
Volume of Communication, Network Size, Identification of Clusters and Subnets, Importance of individual actors (activity, connectedness, brokering position), Identification of isolated/weakly connected actors or subnets, strength/symmetryof relationshipsCumulative or time-window based longitudinal development, growth of thenetwork, frequence analysis, peaks, identification of time-based patterns
Topic-similarity or dissimilarity, cluster formation, categorization, identification of relevant topics, relate actors and links to the overall enterprise contextDissemination of information through the network, finding influencing factors fortime-based patterns by analyzing collaboration- and communication contents and corporate contexts
6) Execute and Implement Actions
5) Deriving Concept and Action Plan
Source: Trier/Bobrik(2007): Systemanalyse und Wissensmanagement. In: Krallmann et al. (2007): Systemanalyse im Unternehmen, in press.
Method for IT-supported Dynamic Network Analysis
© Net Business Center / 10©2007 IKM, SYSEDV, TU Berlin, / 10
The evolution of the most central author’s position in the network
a) 07/01/2000 b) 02/26/2001 Time
c) 10/24/2001 d) 06/21/2002 Time
degree centrality
© Net Business Center / 11©2007 IKM, SYSEDV, TU Berlin, / 11
Evolution of the most central author’s position in his ego-network
a) 02/05/2002 b) 02/07/2002
EGO Vice President Employee Manager President Trader CEO Lawyer
© Net Business Center / 12©2007 IKM, SYSEDV, TU Berlin, / 12
The Application
also see: www.commetrix.de
© Net Business Center / 13©2007 IKM, SYSEDV, TU Berlin, / 13
Measuring the level of network change (as ∆activity)
0
2
4
6
8
10
12
14
Sep 99
Nov 99
Jan 00
Mrz 00
Mai 00
Jul 00
Sep 00
Nov 00
Jan 01
Mrz 01
Mai 01
Jul 01
Sep 01
Nov 01
Jan 02
Mrz 02
Mai 02
Average Relationship Strength
Linear (Average Relationship Strength)
0
50
100
150
200
250
300
350
400
450
500
Sep-99
Nov-99
Jan-00
Mar-00
May-00
Jul-00
Sep-00
Nov-00
Jan-01
Mar-01
May-01
Jul-01
Sep-01
Nov-01
Jan-02
Mar-02
May-02
Number of Active Authors
Number of Active Relations
Time Window = 1mth
© Net Business Center / 14©2007 IKM, SYSEDV, TU Berlin, / 14
Example: Visual Trace of Brokering Actions in the Enron Discourse
a) 02/27/2000 b)
c) d) 02/28/2000
BrokeringActivity
BetweennessCentrality
BrokeringActivity
BetweennessCentrality
1Source: Trier/Bobrik(2007): Analyzing the Dynamics of Community Formation using Brokering Activities. Proceedings CCT 2007. Springer, in press.
© Net Business Center / 15©2007 IKM, SYSEDV, TU Berlin, / 15
That‘s it! Thanks for your attention!
Next issues:
Substantiating dynamic analysis methodsExtending topic analysisDeveloping complex evaluation indicators and visualization ideas!
Collaborations on new network sources and application fields (e.g. social software, corporate applications, virtual communication of online learners) Finding complementary partners for jointresearch funding applications?
Please contact:
Dr.-Ing. Matthias [email protected]
Technical University Berlin, GermanyInstitute for Business Informatics, http://www.sysedv.tu-berlin.de
www.commetrix.deDevelopment Team:
Annette BobrikMatthias Frank Tilmann Bartels
Andreas HoffmannDaniel Mueller
Matthias WerlitzMarkus WallaDanny Graef
Nicola GalanovNancy Kiesel
u.a.