social network analysis course 2010 - 2011
DESCRIPTION
the 5th social media course at skema business school about social network analysisTRANSCRIPT
Social Network AnalysisGuillaume ERETEO
Social Networks
A social network is made of actors (people, organizations, group) that are tied by social links.
Sociograms and graphsactors are represented by points (nodes, vertices) and relations by lines (edges, links, arcs)
• explicit and declared relations
• interactions between actors
• affiliation between actors
Social links
colleague
communicate
like
likeweb
undirected networkswell suited for representing network with symmetric relationships such as facebook
directed networkswell suited for representing network with non reciprocal relationships such as Twitter
weighted networks
well suited for representing the intensity of relationships, the number of interactions (e.g. mails), or the number of affiliations (e.g. shared links).
1
0.5
1
2
3
labelled networkswell suited for representing the type of relationships
family
colleague
father
friend
follow
follow
father
Social network analysis
helps understanding and exploiting the key features of social networks in order to manage
their assets, their life cycle and predict their evolution.
What for?
• To control information flow
• To foster communication
• To improve network resilience
• To trust or not
Diameterthe maximum distance between two actors
Density of links indicates the cohesion of the network
community detection
distribution of actors and activities
community membership
• influences the way information is shared• influences the way actors behave
Centrality: strategic positions
degree centrality focuses on the number of relations:reveal local popularity
[Freeman 1979]
Centrality: strategic positions
closeness centrality focuses on the proximity of actors:reveal capacity to communicate
[Freeman 1979]
Centrality: strategic positionsbeetweenness centrality focuses on intermediary actors reveals brokers and privileged actor in the information flow
[Freeman 1979]
"A place for good ideas"[Burt 1992] [Burt 2004]
Centrality: strategic positionseigen vector centrality focuses on the connections of neighboursreveals actors that are well connected
transitivity of relationshipsevaluating social opportunities
Peter Jack
Paul
tendency to bind with similar others
"Birds of a feather flock together"
"interaction produces similarity, while similarity produces interaction"
[Mika 2005]
Preferential attachement
who will increase its popularity?
Balance Theory[Heider 1958]
Critical mass
Many ways of connecting people!
likes
ingredient
typemainDish
Food
subclassOf
type
Who has the best betweeness?this is a matter of semantic
knows in passing
has met
colleague
works With
works With
friend
Metrics are not sufficient enough
knows in passing
has met
colleague
works With
works With
friend
examples of vizualisation and analysis tools
Touchgraph facebook browser
easily vizualize your facebook social network
http://apps.facebook.com/touchgraph
this is mine
Analyze your facebook network
1. Extract it first with netvizz: http://apps.facebook.com/netvizz
2. Open it with my favorite graph visualization tool: Gephi http://www.gephi.org
extract with Netvizz
• choose the information to include
• Save the file
analyze with Gephi
Your network is undirected
select the force layout
detect communities
highlight communities
compute shortest paths of relations
highlight most popular actors
Highlight brokers
print attributes
NODEXL
• Excel 2007 based social network visualization and analysis tool
• Homepage http://nodexl.codeplex.com
• Tutorial http://casci.umd.edu/images/4/46/NodeXL_tutorial_draft.pdf
Quick overview of the tutorial
"One way to begin using NodeXL is to type in your own edge list. For example, you might type the name of people who are friends in each row filling in the Vertex 1 and Vertex 2 columns"
"Click on the Show Graph button (directly above the graph pane) to show the network of friendships"
"To calculate graph metrics first click on the Graph Metric button on the Analysis section of
the NodeXL Ribbon."
"Vertex metrics can be mapped onto visual attributes. The graph legend shows that Degree is mapped to Size
and Betweenness Centrality is mapped to Opacity."
You can now handle your social capital
The social capital is the "resources embedded in one's social networks, resources that can be accessed or mobilized through ties in the networks" [Lin 2008]
http://www.kstoolkit.org/Social+Network+Analysis
nameGuillaume Erétéo
account
mailmentorOf
mentorOf
organizationorganization
manages
contributes
contributes answers
twitter.com/ereteog