Jure Leskovec Joint work with Eric Horvitz, Microsoft Research.

Slides:



Advertisements
Similar presentations
Routing in Poisson small-world networks A. J. Ganesh Microsoft Research, Cambridge Joint work with Moez Draief.
Advertisements

Mobile Communication Networks Vahid Mirjalili Department of Mechanical Engineering Department of Biochemistry & Molecular Biology.
Stelios Lelis UAegean, FME: Special Lecture Social Media & Social Networks (SM&SN)
Λ14 Διαδικτυακά Κοινωνικά Δίκτυα και Μέσα Strong and Weak Ties Chapter 3, from D. Easley and J. Kleinberg book.
Online Social Networks and Media Navigation in a small world.
Advanced Topics in Data Mining Special focus: Social Networks.
1 Evolution of Networks Notes from Lectures of J.Mendes CNR, Pisa, Italy, December 2007 Eva Jaho Advanced Networking Research Group National and Kapodistrian.
Emergence of Scaling in Random Networks Barabasi & Albert Science, 1999 Routing map of the internet
Network Models Social Media Mining. 2 Measures and Metrics 2 Social Media Mining Network Models Why should I use network models? In may 2011, Facebook.
UNDERSTANDING VISIBLE AND LATENT INTERACTIONS IN ONLINE SOCIAL NETWORK Presented by: Nisha Ranga Under guidance of : Prof. Augustin Chaintreau.
CS 728 Lecture 4 It’s a Small World on the Web. Small World Networks It is a ‘small world’ after all –Billions of people on Earth, yet every pair separated.
Marko Grobelnik, Dunja Mladenic JSI Parts of the presentation taken from the tutorial “Structure and function of real-world graphs and networks” by Jure.
Graphs and Topology Yao Zhao. Background of Graph A graph is a pair G =(V,E) –Undirected graph and directed graph –Weighted graph and unweighted graph.
1 Algorithms for Large Data Sets Ziv Bar-Yossef Lecture 7 May 14, 2006
A Measurement-driven Analysis of Information Propagation in the Flickr Social Network WWW09 报告人: 徐波.
Mining Large Graphs Part 3: Case studies Jure Leskovec and Christos Faloutsos Machine Learning Department Joint work with: Lada Adamic, Deepay Chakrabarti,
Large-scale organization of metabolic networks Jeong et al. CS 466 Saurabh Sinha.
(Social) Networks Analysis III Prof. Dr. Daning Hu Department of Informatics University of Zurich Oct 16th, 2012.
Topic 13 Network Models Credits: C. Faloutsos and J. Leskovec Tutorial
Planetary-Scale Views on a Large Instant-Messaging Network 报告人:徐波.
Jure Leskovec Joint work with Eric Horvitz, Microsoft Research.
Jure Leskovec, CMU Eric Horwitz, Microsoft Research.
Jure Leskovec Computer Science Department Cornell University / Stanford University Joint work with: Eric Horvitz, Michael Mahoney,
Section 8 – Ec1818 Jeremy Barofsky March 31 st and April 1 st, 2010.
Network properties Slides are modified from Networks: Theory and Application by Lada Adamic.
Web Science Course Lecture: Social Networks - * Dr. Stefan Siersdorfer 1 * Figures from Easley and Kleinberg 2010 (
Small World Social Networks With slides from Jon Kleinberg, David Liben-Nowell, and Daniel Bilar.
Network Characterization via Random Walks B. Ribeiro, D. Towsley UMass-Amherst.
Internet Economics כלכלת האינטרנט Class 10 – it’s a small world 1.
1 Announcing … Global broadband subscribers to 30 June 2005 Total: 176 million 115 million * 65% * choose DSL.
Jure Leskovec Computer Science Department Cornell University / Stanford University Joint work with: Jon Kleinberg (Cornell), Christos.
Microsoft Instant Messenger Communication Network How does the world communicate? Jure Leskovec Machine Learning Department
Online Social Networks and Media
Λ14 Διαδικτυακά Κοινωνικά Δίκτυα και Μέσα Networks and Surrounding Contexts Chapter 4, from D. Easley and J. Kleinberg book.
3. SMALL WORLDS The Watts-Strogatz model. Watts-Strogatz, Nature 1998 Small world: the average shortest path length in a real network is small Six degrees.
CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University
Complex Network Theory – An Introduction Niloy Ganguly.
Women in physics : Are we there yet? 1.  Statistics from ATLAS and CERN  Are women physicists treated equally?  Easy things to improve the situation.
Lecture 10: Network models CS 765: Complex Networks Slides are modified from Networks: Theory and Application by Lada Adamic.
Social Network Analysis. Outline l Background of social networks –Definition, examples and properties l Data in social networks –Data creation, flow and.
Complex Network Theory – An Introduction Niloy Ganguly.
Most of contents are provided by the website Network Models TJTSD66: Advanced Topics in Social Media (Social.
What Is A Network? (and why do we care?). An Introduction to Network Theory | Kyle Findlay | SAMRA 2010 | 2 “A collection of objects (nodes) connected.
Tools for large graph mining WWW 2008 tutorial Part 4: Case studies Jure Leskovec and Christos Faloutsos Machine Learning Department Joint work with: Lada.
1 Friends and Neighbors on the Web Presentation for Web Information Retrieval Bruno Lepri.
Chapter 4 Histograms Stem-and-Leaf Dot Plots Measures of Central Tendency Measures of Variation Measures of Position.
Small World Social Networks With slides from Jon Kleinberg, David Liben-Nowell, and Daniel Bilar.
1 Patterns of Cascading Behavior in Large Blog Graphs Jure Leskoves, Mary McGlohon, Christos Faloutsos, Natalie Glance, Matthew Hurst SDM 2007 Date:2008/8/21.
Offense: Planetary-Scale Views on a Large Instant Messaging Network J. Leskovec, et al.
Class 4: It’s a Small World After All Network Science: Small World February 2012 Dr. Baruch Barzel.
Netlogo demo. Complexity and Networks Melanie Mitchell Portland State University and Santa Fe Institute.
Topics In Social Computing (67810) Module 1 Introduction & The Structure of Social Networks.
Chapter 4 Histograms Stem-and-Leaf Dot Plots Measures of Central Tendency Measures of Variation Measures of Position.
1© GfK 2016 | Concerns about safety and security Concerns about safety and security Global GfK survey July 2016.
Social Networks Some content from Ding-Zhu Du, Lada Adamic, and Eytan Adar.
Lecture 23: Structure of Networks
Groups of vertices and Core-periphery structure
The Role of Optimal Distinctiveness and Homophily in Online Dating
How Do “Real” Networks Look?
Lecture 23: Structure of Networks
The Watts-Strogatz model
How Do “Real” Networks Look?
How Do “Real” Networks Look?
How Do “Real” Networks Look?
Network Science: A Short Introduction i3 Workshop
Lecture 23: Structure of Networks
CS224w: Social and Information Network Analysis
Lecture 9: Network models CS 765: Complex Networks
Advanced Topics in Data Mining Special focus: Social Networks
Planetary-Scale Views on a Large Instant-Messaging Network
Presentation transcript:

Jure Leskovec Joint work with Eric Horvitz, Microsoft Research

 Contact (buddy) list  Messaging window 2

3 Buddy Conversation

 Observe social and communication phenomena at a planetary scale  Largest social network analyzed to date 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? 4

 For every conversation (session) a list of participants:  User Id  Time Joined  Time Left  Number of Messages Sent  Number of Messages Received  There can be multiple participants per conversation  Everything is anonymized. No message text 5

 User demographic data (self-reported):  Age  Gender  Location (Country, ZIP)  Language 6

 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 exchanged messages 7

Activity on a typical day (June ):  1 billion conversations  93 million users login  65 million different users talk (exchange messages)  1.5 million invitations for new accounts sent 8

How does user demographics influence communication? 9

10

11 People tend to talk to similar people (except gender)  How do people’s attributes (age, gender) influence communication? Probability that users share an attribute

12 User self reported age High Low 1) Young people communicate with same age 2) Older people communicate uniformly across ages

13 User self reported age High Low 1) Old people talk long 2) Working ages (25-40) talk short

14 User self reported age High Low 1) Old people talk long 2) Working ages (25-40) talk quick

15 User self reported age High Low 1) Old people talk slow 2) Young talk fast

 Is gender communication biased?  Homophily: Do female talk more among themselves?  Heterophily: Do male-female conversations took longer?  Findings:  Num. of. conversations is not biased (follows chance)  Cross-gender conversations take longer and are more intense (more attention invested) 16 MF 49% 21% 20% Conversations MF 5 min 4.5 min 4min Duration MF Messages/conversation

 Longer links are used more 17

Each dot represents number of users at geo location 18 Map of the world appears! Costal regions dominate

Users per capita 19 Fraction of country’s population on MSN: Iceland: 35% Spain: 28% Netherlands, Canada, Sweden, Norway: 26% France, UK: 18% USA, Brazil: 8%

20 Digital darkness, “Digital Divide”

21 For each conversation between geo points (A,B) we increase the intensity on the line between A and B

22

23

 Max number of people simultaneously talking is 20, but conversation can have more people 24

Sessions between fewer people run out of steam 25

26 At least 1 message exchanged

 Buddy graph  240 million people (people that login in June ’06)  9.1 billion buddy edges (friendship links)  Communication graph (take only 2-user conversations)  Edge if the users exchanged at least 1 message  180 million people  1.3 billion edges  30 billion conversations 27

28 Number of buddies follows power- law with exponential cutoff distribution Limit of 600 buddies

29 There is “no average” degree. But degrees are heavily skewed. “Heavy tailed” or “power law” distributions

30

Milgram’s small world experiment (i.e., hops + 1) 31  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 edges 6 degrees of separation

32 MSN Messenger network Number of steps between pairs of people Avg. path length % of the people can be reached in < 8 hops HopsNodes ,96 48,648 53,299, ,395, ,059, ,995, ,321, ,955, , , , , , ,

 Degrees of separation (avg. shortest path length) inside the country 33 CountyCountry Avg Path Len [hops] Turkey 5.18 Brazil 5.60 Belgium 5.63 United Kingdom 5.63 Spain 5.72 Mexico 5.72 France 6.03 China 6.38 United States 6.96

34 CountyCountry Avg Path Len [hops] United StatesLebanon6.17 United StatesAustralia6.22 United StatesNorway6.23 United StatesAlbania6.24 United StatesMalta6.24 United StatesUnited Kingdom6.28 United StatesBahamas6.29 United StatesSweden6.37 United StatesBahrain6.37 United StatesCanada6.38 CountyCountry Avg Path Len [hops] United StatesBulgaria7.28 United StatesPoland7.39 United StatesRussia7.42 United StatesRomania7.48 United StatesLithuania7.57 United StatesSlovakia7.84 United StatesKorea, South8.03 United StatesCzech Republic8.05 United StatesJapan8.85 Top “close” countries Top “far” countries

 How many triangles are closed?  Clustering normally decays as k High clusteringLow clustering Communication network is highly clustered: k -0.37

 What is the structure of the core of the network? 36 [Batagelj & Zaveršnik, 2002]

 People with k<20 are the periphery  Core is composed of 79 people, each having 68 edges among them 37

 Remove nodes (in some order) and observe how network falls apart:  Number of edges deleted  Size of largest connected component 38 Order nodes by:  Number of links  Total conversations  Total conv. Duration  Messages/conversation  Avg. sent, avg. duration

39

40

 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 41

 J. Leskovec and E. Horvitz: Worldwide Buzz: Planetary-Scale Views on an Instant- Messaging Network, WWW 2008 