Network Architectures & Services
Fernando Kuipers (
[email protected]
)
Network
“hardware”
Network
“software”
Network
“peopleware”
Individual: Quality of Experience
Friends: Recommendation
Global: Online Social Networks
Processes: Viruses, tweets, …
Algorithms: Routing, SDN, …
Big data: Mining & analysis
Complex networks: Internet, energy,
brain, optical networks, …
Network design: Robustness vs. cost
M
ult
i-d
im
ens
io
na
l a
na
lys
is
Not only Happiness is Contagious
Online Social Network Analysis in the Context of a Changing
Communication Paradigm in Telecommunications
Norbert Blenn
The Change of Communication Paradigms
“The challenge is not just in understanding the
technology, but also the fundamental shifts in human
communication behavior.”
(IBM Institute for Business Value analysis)
Mark
et d
irec
tion
Mark
et d
irec
tion
Mark
et d
irec
tion
point to point
conversational
many to many
collaborative
Mark
et d
irec
tion
Mark
et d
irec
tion
Mark
et d
irec
tion
Mark
et d
irec
tion
Mark
et d
irec
tion
Mark
et d
irec
tion
point to point
conversational
many to many
collaborative
(TeleGeography Report)
Year
2005
2006
2008
2010
2012
Skype’s international 2.9 %
4.4 %
8 %
13 %
34 %
call market share
Effects of the Change of
provider-controlled
Open Internet
traditional
communication
shared social
space
Effects of the Change of
Telecommunication and
Online Social Network Analysis
How to benefit from the change in communication?
•
Big data analysis and the combination of Big Data to (Big Data)
n
in order to create new services and value cases
•
New services through mobility analysis
•
Real-time feedback through social sensors
•
Product or company placement in the social media landscape
•
Recommendation systems
Traces of two users from Eindhoven
Work
Shopping
Restaurants
Living
Users of Online Social Networks as
“Social Sensors”
Mobility-Networks
Identifying regions in which individuals travel/use
mobile servives enables location based services
and valuable results for other economic sectors.
Users of Online Social Networks as
“Social Sensors”
Real-time feedback
of product/company placement
"Context-Sensitive Sentiment Classification of Short Colloquial Text", N. Blenn, C. Doerr, K. Charalampidou and P. Van Mieghem, IFIP Networking 2012
KPN
Vodafone
Users of Online Social Networks as
“Social Sensors”
Combination of Big Data into (Big Data)
n
Sentiment analysis of tweets
negative
“Word-Of-Mouth” in Online Social Networks
Influential users:
are more “central”,
have high “betweenness”,
having a higher “effectivity”
Contagious opinions and recommendations
"Lognormal Infection Times of Online Information Spread", C. Doerr , N. Blenn, and P. Van Mieghem, 2013, PLoS ONE
“Word-Of-Mouth” in Online Social Networks
•
The more friends forward a message, the more likely the friend will adopt
Contagious opinions
Understanding Interests of Individuals
Recommendation systems
"How much do your friends know about you? Reconstructing private information from the friendship graph ", N. Blenn, C. Doerr, N. Shadravan and P. Van Mieghem, Eurosys 2012, 5th Workshop on Social Network Systems
•
Knowing more about the user than he/she himself
•
Our Datasets / Big Data
•
Twitter: 3 billion Messages (ca. 200 new per second),
6.1 billion edges (friendship relations), 120 million profiles
•
Hyves: 5.8 million profiles, 90 million edges (friendship relations)
•
IMDB: 178,000 movies, 2million comments
(complete)
•
Digg: 2 million profiles, 7.7 million edges, 315 million votes
(complete)
•
Sourceforge: 100,000 projects, 460,000 user
(complete)
•
…
Telecommunication and
Online Social Network Analysis
Reliable information because of “(Big Data)
n
”
What did he just say?
•
Understanding the shift in people’s
communication pattern states a high potential
to create new untapped value
•
Knowing how customers behave and how
they interact will be the key driver in the
future to remain competitive
•
Network and content analysis can predict
trends as … (Influential users and groups can
be found when approached the right way).
•
“(Big Data)
n
” is
the
potential for the telecom
industry as no one else except companies in
the telecom sector have access to all
necessary datasets
Thinking out of the Box
You found the last slide.
References:
• "Crawling and Detecting Community Structure in Online Social Networks using Local Information", N. Blenn, C. Doerr, S. van Kester and P. Van Mieghem, 2012, IFIP Networking 2012, May 21-25, Prague, Czech Republic.
• "Metric Convergence in Social Network Sampling", C. Doerr and N. Blenn, 2013,
SIGCOMM 2013, the 5th ACM HotPlanet Workshop
• "Lognormal infection Times of online information spread", C. Doerr , N. Blenn, and P. Van Mieghem, 2013, PLoS ONE (to appear)
• "Are Friends Overrated? A Study for the Social Aggregator Digg.com", C. Doerr, N. Blenn, S. Tang and P. Van Mieghem, Computer Communications, 35(7), pp. 796--809, 2012. DOI 10.1016/j.comcom.2012.02.001, 2012
• "Context-Sensitive Sentiment Classification of Short Colloquial Text", N. Blenn, C. Doerr, K. Charalampidou and P. Van Mieghem, 2012, IFIP Networking 2012, May 21-25, Prague, Czech Republic.
• "Digging in the Digg Social News Website"; S. Tang, N. Blenn, C. Doerr and P. Van Mieghem, 2011; IEEE Transactions on Multimedia, Vol. 13, No. 5, October, pp. 1163-1175.
• "How much do your friends know about you? Reconstructing private information from
the friendship graph ", N. Blenn, C. Doerr, N. Shadravan and P. Van Mieghem, 2012, Eurosys 2012, 5th Workshop on Social Network Systems
• "Lognormal Distribution in the Digg Online Social Network"; P. Van Mieghem, N. Blenn and C. Doerr, 2011, The European Physical Journal B, Vol. 83, No. 2, pp. 251-261.
• "Content Propagation in Online Social Networks"; N. Blenn, C. Doerr, P. Van Mieghem, ICTOpen 2011
• "Characterizing the Structure of Affiliation Networks", D. Liu, N. Blenn and P. Van Mieghem, 2012, 12th International Conference on Computational Science (ICCS), June 4-6, Omaha, Nebraska, USA.
• “A Social Network Model Exhibiting Tunable Overlapping Community Structure", D. Liu, N. Blenn and P. Van Mieghem, 2012,1st International Workshop on Advances in Computational Social Science, June 4-6, Omaha, Nebraska, USA.