Jure Leskovec, CMU Eric Horwitz, Microsoft Research.
-
Upload
francis-george -
Category
Documents
-
view
234 -
download
1
Transcript of Jure Leskovec, CMU Eric Horwitz, Microsoft Research.
![Page 1: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/1.jpg)
Planetary Scale Views on an Instant Messenger Communication Network
Jure Leskovec, CMUEric Horwitz, Microsoft Research
![Page 2: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/2.jpg)
2
Instant Messaging
Contact (buddy) list Messaging window
![Page 3: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/3.jpg)
3
Instant Messaging as a Network
Buddy Conversation
![Page 4: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/4.jpg)
4
IM – Phenomena at planetary scale
Observe social and communication phenomena at planetary scale
Practically the largest social network
Research questions: How does communication change with user
demographics (age, sex, language, country)? How does geography affect communication? What is the structure of the communication
network?
![Page 5: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/5.jpg)
5
Communication data
The record of communication User demographic data (self-reported):
Age Gender Location (Country, ZIP) Language
Communication data: For each conversation: participants and time No message text
![Page 6: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/6.jpg)
6
Data statistics: Total activity We collected the data for June 2006 Log size:
150Gb/day (compressed) Total: 1 month of communication data:
4.5Tb of compressed data Activity over June 2006 (30 days)
245 million users logged in 180 million users engaged in conversations 17,5 million new accounts activated More than 30 billion conversations More than 255 billion messages
![Page 7: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/7.jpg)
7
Data statistics: Typical day
Activity on June 1 2006: 1 billion conversations 93 million users login 65 million different users talk (exchange
messages) 1.5 million invitations for new accounts sent
![Page 8: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/8.jpg)
Part 3-8
User & Communication characteristics
How does user demographics influence communication?
![Page 9: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/9.jpg)
9
User characteristics: Age
![Page 10: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/10.jpg)
10
Age piramid: MSN vs. the world
![Page 11: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/11.jpg)
Part 3-11
Communication: Demographics
Correlation
People tend to talk to similar people (except gender)
How do people’s attributes (age, gender) influence communication?
Probability
![Page 12: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/12.jpg)
12
Age: Number of conversations
Use
r se
lf r
eport
ed a
ge
High
Low
1) Young people communicate with same
age2) Older people
communicate uniformly across ages
![Page 13: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/13.jpg)
13
Age: Total conversation duration
Use
r se
lf r
eport
ed a
ge
High
Low
1) Old people talk long2) Working ages (25-40)
talk short
![Page 14: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/14.jpg)
14
Age: Messages per conversation
Use
r se
lf r
eport
ed a
ge
High
Low
1) Old people talk long2) Working ages (25-40)
talk quick
![Page 15: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/15.jpg)
15
Age: Messages per unit time
Use
r se
lf r
eport
ed a
ge
High
Low
1) Old people talk slow2) Young talk fast
![Page 16: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/16.jpg)
16
Communication: Gender
Is gender communication biased? Do female talk more among themselves? Do male-female conversations took longer?
Findings: No of. conversations is not biased (follows chance) Cross gender conversations take longer and are mo
intense (people put more attention)
M F49%21%20%
Conversations
M F5mi n4.5 min4min
Duration
M F7.66.65.9
Messages/conversation
![Page 17: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/17.jpg)
17
Communication: Geo distance
Longer links are used more
![Page 18: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/18.jpg)
18
Communication: Geography (1)
Each dot represents number of users at geo locationMap of the world appears!Costal regions dominate
![Page 19: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/19.jpg)
19
Communication: Geography (2)
Users per capita
Mid-USA and Scandinavia have highest per capita
Fraction of country’s population on MSN:•Iceland: 35%•Spain: 28%•Netherlands, Canada, Sweden, Norway: 26%•France, UK: 18%•USA, Brazil: 8%
![Page 20: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/20.jpg)
20
World communication axis
Northern hemisphere dominates communication
For each conversation between geo points (A,B) we increase the intensity on the line between A and B
![Page 21: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/21.jpg)
Messaging as a Network
21At least 1 message exchanged
![Page 22: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/22.jpg)
22
IM Communication Network Buddy graph (estimate)
240 million people (people that login in June ’06) 9.1 billion buddy edges (friendship links)
Communication graph Edge if the users exchanged at least 1 message 180 million people 1.3 billion edges 30 billion conversations
![Page 23: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/23.jpg)
23
Network: Number of buddies
Number of buddies follows power-law with exponential
cutoff distribution
Limit of 600 buddies
![Page 24: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/24.jpg)
24
Network: Small world
Milgram’s small world experiment
(i.e., hops + 1)
Small-world experiment [Milgram ‘67] People send letters from Nebraska to Boston
How many steps does it take? Messenger social network of the whole planet Eart
240M people, 1.3B edgesMSN Messenger network
Number of steps
between pairs of people
Avg. path length 6.690% of the people can be reached in <
8 hops
![Page 25: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/25.jpg)
25
Network: Clustering
How many triangles are closed?
Clustering normally decays as k-1
High clustering Low clustering
Communication network is highly clustered: k-0.37
![Page 26: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/26.jpg)
26
Network: Connectivity
![Page 27: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/27.jpg)
27
Network: k-Cores decomposition
What is the structure of the core of the network?
[Batagelj & Zaveršnik, 2002]
![Page 28: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/28.jpg)
28
k-Cores: core of the network
People with k<20 are the periphery Core is composed of 79 people, each having 68
edges among them
![Page 29: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/29.jpg)
29
Network: Tie-strength
Remove nodes (in some order) and observe how network falls apart: Number of edges deleted Size of largest connected component
Order nodes by: Number of links Total conversations Total conv. Duration Messages/conversation Avg. sent, avg. duration
![Page 30: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/30.jpg)
30
Strength: Nodes vs. Edges
![Page 31: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/31.jpg)
31
Strength: Connectivity
![Page 32: Jure Leskovec, CMU Eric Horwitz, Microsoft Research.](https://reader036.fdocuments.net/reader036/viewer/2022062304/56649e5f5503460f94b58d1d/html5/thumbnails/32.jpg)
32
Conclusion
Social network of the whole planet Earth The largest social network analyzed
Strong presence of homophily people that communicate are similar (except gender)
Well connected Small-world in only few hops one can research most of the
network Very robust:
many (random) people can be removed and the network is still connected