2 nd IEEE International Conference on Social Computing

  • Slides: 23
Download presentation
2 nd IEEE International Conference on Social Computing Link creation and profile alignment in

2 nd IEEE International Conference on Social Computing Link creation and profile alignment in the a. Nobii social network Speaker: Luca Maria Aiello, Ph. D student aiello@di. unito. it Authors: Università degli Studi di Torino ISI Foundation Luca Maria Aiello Giancarlo Ruffo Rossano Schifanella Alain Barrat Ciro Cattuto Keywords : link creation, homophily, social influence, a. Nobii

Open questions in social network analysis What are the dynamics leading to link creation?

Open questions in social network analysis What are the dynamics leading to link creation? 2. What is the interplay between user similarity and link creation? 1. 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 2

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 3

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 4

Social network for bookworms �Data-driven analysis on anobii. com �Profile features �Social network ◦

Social network for bookworms �Data-driven analysis on anobii. com �Profile features �Social network ◦ Library and wishlist ◦ Groups ◦ Tags 4 th snapshot ◦ Directed ◦ Friendship + neighborhood Friendship Neighborhood Union Nodes 74, 908 54, 590 86, 800 Links 268, 655 429, 482 697, 910 � 6 snapshots, 15 days apart �Full giant connected component 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 5

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 6

Basic statistics <kout> 8. 0 Reciprocation 0. 57 Avg SPL 5. 3 Diameter 20

Basic statistics <kout> 8. 0 Reciprocation 0. 57 Avg SPL 5. 3 Diameter 20 �Broad distributions �High reciprocation �High diameter 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 7

Correlations and mixing patterns Pearson correlation kout ng nb nw 1 0. 31 0.

Correlations and mixing patterns Pearson correlation kout ng nb nw 1 0. 31 0. 18 1 0. 32 0. 31 1 0. 22 � Positive correlations between: � Connectivity and activity � Different activities � Assortativity (n. s. ) 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 8

Profile similarity vs. social distance Does similarity between user profiles depend on the social

Profile similarity vs. social distance Does similarity between user profiles depend on the social distance? �Topical overlap �Statistical correlation because of assortative biases? �Null model to discern real overlap from purely statistical effects ◦ No topical overlap other than that caused by statistical mixing patters 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 9

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 10

Motivations �…does geographical overlap hold in the network as well? �Dataset peculiarities ◦ Many

Motivations �…does geographical overlap hold in the network as well? �Dataset peculiarities ◦ Many users specify their home country (97%) or town (38%) ◦ Particular community distribution 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 11

Geographical clustering Country-level social network Zoom on Italy 22/08/2010 Social. Com 2010 - Luca

Geographical clustering Country-level social network Zoom on Italy 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 12

Geographic and language overlap � Null model test with random link rewire � Country-level

Geographic and language overlap � Null model test with random link rewire � Country-level overlap due to language barriers � City-level overlap for friendship (trivial…) � City-level overlap for neighborhood ◦ Bidirectional causality connection between acquaintance in real life and connectivity in the online social network 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 13

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 14

Triadic closure �Classification of new links at time t+1 between nodes already present at

Triadic closure �Classification of new links at time t+1 between nodes already present at time t (t ∈ {1, …, 5}) Closure Direct Reciprocated 75% 20% Bidirectional 30% 25% Double closure 10% �Reciprocation is strong �Users tend to choose “friends of their friends” as new friends 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 15

Proximity-driven attachment � Users tend to choose “friends of their friends” or people close

Proximity-driven attachment � Users tend to choose “friends of their friends” or people close in the social network as new friends � This process results in preferential attachment 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 16

Causality between similarity and link creation �Topical overlap is observed for all profile features

Causality between similarity and link creation �Topical overlap is observed for all profile features What is the cause of topical overlap? �Three possible explanations: 1. Homophily (people connect with similar people) 2. Social influence (social connection conveys similarity) 3. Mixture of the two �Explore the causality relationship between profile similarity and social linking 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 17

Similarity link creation duv = 2 u→v u↔v �n cb� 9. 5 12. 9

Similarity link creation duv = 2 u→v u↔v �n cb� 9. 5 12. 9 18. 5 σb 0. 02 0. 04 �n cg� 1. 12 1. 10 1. 67 σg 0. 05 0. 08 0. 11 Closure Dbl closure 18. 2 23. 4 0. 05 1. 81 1. 20 0. 12 � Average similarity of pairs forming new links between t 0 and t 0+1 (t 0=4), compared with average similarity of all the pairs at distance 2 at time t 0 � Pairs that are going to get connected show a substantially higher similarity 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 18

Books Groups Link creation similarity �Evolution of the similarity between pairs linking together at

Books Groups Link creation similarity �Evolution of the similarity between pairs linking together at different times 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 19

 • Dataset • Static analysis s i s y l a n a

• Dataset • Static analysis s i s y l a n a c i h p a • Geogr sis y l a n a l a c i m a n y • D • Conclusions 22/08/2010 Outline Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 20

Summary � What are the dynamics that rule link creation? ◦ Reciprocation (in direct

Summary � What are the dynamics that rule link creation? ◦ Reciprocation (in direct networks) ◦ Triadic closure ◦ Proximity-driven (preferential) attachment �On geographical space �On the social network ◦ Language-driven attachment ◦ Homophily What is the interplay between user similarity and link creation? � ◦ Tight coupling (topical overlap) ◦ Topical overlap is caused by homophily and social influence both 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 21

Future work �Link prediction �Information spreading �Extend analysis to other social systems 22/08/2010 Social.

Future work �Link prediction �Information spreading �Extend analysis to other social systems 22/08/2010 Social. Com 2010 - Luca Maria Aiello, Università degli Studi di Torino 22

2 nd IEEE International Conference on Social Computing Thank you for your attention! Speaker:

2 nd IEEE International Conference on Social Computing Thank you for your attention! Speaker: Luca Maria Aiello aiello@di. unito. it www. di. unito. it/~aiello