从社交网络到移动社交网络

Nokia Research Center

From Social Networks to Mobile Social Networks

Zhen Liu

Lab Director

Nokia Research Center

Dec. 9, 2011

•Social networking landscape •Mobile social networking •State-of-the-art •Ephemeral social networks •Our research

•Summary

Strong tie

Generic

Specific Social Networking Matrix Divided

by Tie Strength & Scope

Source: http://www.internetworldstats.comSocial Media on Mobile in the U.S.

Source: Nielsen http://cn.nielsen.com/documents/SocialMediaReport_SCN.pdf, Q3 2011Social Networking on Mobile in China

Most Downloaded iPhone Free Apps

Source: 同步推http://app.tongbu.com/iphone/paihang/itunes-cn-6005-free, Nov. 2011

Is this really mobile social networking? No, this is social networking on mobile!

Mobile Social Networking

Making the mobile an integral part of your social network and lifestyle distributed content sharing

social networks

sensor networks

pervasive computing

Integrate contact, content

and context into mobile

social networks

Provide integrated experience by fusing physical and digital

social interactions through the mobileSocial Networking on Mobile vs. Mobile Social Networking

Feature Social networking

on mobile

Mobile social

networking

Duration Continuous Ephemeral Offline interaction None Activity-based Proximity and Location Co-location Nearby, encounter

Context sensing None Environment and phone

sensing

Contact management Indirect discovery Direct discovery Content sharing Public Public, Private, Group Collaboration Coordination Ad-hocIndustry Efforts in Mobile Social Networking

Contact

Content Context•Opportunistic networking

•Familiar Stranger (Paulos and Goodman, 2002)

•Cityware(Kostakos, 2008)

•Conferator(Atzmueller et al, 2011)

•User behavior vs. online social networks

•Probability of user joining online community increases as number of friends already there increases (Backstrom et al, 2006)

•Correlation between social connectivity and tag vocabulary (Marlow et al, 2006)

•Relationship between communication and personal behavior, using

demographic data (Singla et al, 2008)

•User behaviors in above are conducted online, or inferred by their

online profiles•Proximity as a social pattern of user behavior

•Using GPS (Eagle and Pentland, 2009),

•Location entropy (Crawnshaw et al, 2010)

•Bluetooth to discover the relative physical closeness (Quercia and Capra, 2009)

•Sensing and phone context (Madan et al, 2010; Aharony et al, 2011)

•Enhancing social interactions at conferences using RFID proximity

(Atzmueller et al, 2011)

Gap between Offline and Online Social Networks

How many of your

friends are really your

friends? (V. Miller, Associated Content,

July 27, 2009)

But Social Networking is Done Mostly Offline!

Contact

ESN = AC 3

Activity

Context

Content

Using ESN to Bridge Online to Offline and vice versa Online social networks (FB, Twitter, Weibo, Renren …)

A c t i v i t y 1

A c t i v i t y 3

A c t i v i t y n Offline physical activities (Conf., Meeting, Party, Shopping, Hiking…)

A c t i v i t y 2

…

time

•In physical world, social networks are built on activities •ESN mirrors your offline memorable moments in online world •ESN enhances your O2O social interactions to help form your social networkResearch Questions: Network and Model

•What are the characteristics of the ESN?

•Network, social, temporal, content, environment •How to collect context to infer social community

intelligence?

•Positioning and proximity, social, collaborative sensing,

behavioural modeling, group and activity identification •How to record, model and identify an ESN?

•What is the interaction between online to offline and vice versa, and how to persist from offline to online?

•Overlap and transition, network models

Research Questions: User and Social

•How to record and manage social relationships?•How to handle privacy and trust in an ESN and how does this change when transition to online?•How to use the ESN for distribution and

recommendation of content, contact and context?What We Have Done

•Developed a platform for ephemeral social

networking called Nokia Find & Connect •Examined characteristics of offline encounter network •Determined how offline encounters influences online connectionsOur Research Questions

What are the characteristics of an ESN?

How to record, model and identify the

ESN?

What is the interaction between offline

and online?Nokia Find & Connect: Platform for Enabling ESN

Find people and rooms Connect with people

Source: Chang et al, Enhancing the Experience and Efficiency at a Conference with Mobile Social Networking: Case Study with Find & Connect, In Proc. of HumanCom2011, 2011Social Networks from Nokia Find & Connect

Property GCJK follow GCJK contacts GCJK encounters

# of users724170

# of links12351592

Average degree

1.7083

2.58.46

Network density

0.024060.0621950.246

Network diameter

6

Average clustering

coefficient

0.2210.1950.683

Average shortest path

length

2.7827 2.6229 2.02

Source: Xu et al, Physical Proximity and Online User Behavior in an Indoor Mobile Social Networking Application, In Proc. of CPSCom2011, 2011Statistical Characteristics of Online Social Networks Property Brightkite Foursquare Twitter LiveJournal

Total nodes54,19058,424409,093992,886

Total links213,668351,216182,986,35329,5,952

4.71 4.60 2.77 4.

Avg. shortest path

length

Network diameter NA NA NA NA Avg. degree7.8812.02447.2929.85

Cluster coeff.0.1810.2560.2070.185

Network density NA NA NA NA

Source: Scellato et al, Distance Matters: Geo-social Metrics for Online Social Networks, In Proc. of 3rd Workshop on Online Social Networks, 2010

Offline Encounters Influences Online Friendship

Source: Xu et al. Social Linking and Physical Proximity in a Mobile Location-based Service, 1st International Workshop on Mobile Location-based Services, In Proc. of UbiComp2011, 2011Offline Improves Friend Recommendation

Source: Xu et al. Using Physical Context in a Mobile Social Networking Application for Improving Friend Recommendations, 1st International Workshop on Sensing, Networking and Computing on Smartphones, In Proc. of CPSCom2011, 2011System Challenges

•Accuracy and relevance of finding ESN

•Proximity and positioning accuracy

•Activity detection and context recognition •Scalability and performance

•Immediate conversion and high transfer rate from

offline to online, as well as persistence from offline to online and vice versa

•…..Summary

•We are moving from social networking to mobile

social networking

•Mobile social networking makes the mobile an

integral part of your social network and lifestyle •Context and social intelligence are needed to drive

mobile social networking forward and make it a daily reality: ephemeral social networks