Microblogging Behavior on Sina Weibo and Twitter

Report
A Comparative Study of Users’
Microblogging Behavior on Sina Weibo
and Twitter
Qi Gao
Web Information Systems
Delft University of Technology
Delft
University of
Technology
What we study: microblogging behavior
340,000,000
500,000,000
100,000,000
250,000,000
What are the differences in Chinese and
Western users’ microblogging behavior?
Wikipedia:Twitter
Wikipedia:Sina_Weibo
2 Microblogging Behavior on Sina Weibo and Twitter
Cultural differences
standing up for himself
Flickr:semicharmed
acting as a member of a group
Flickr:webei
3 Microblogging Behavior on Sina Weibo and Twitter
Data sources
follower/followee
network
limited length of
a post
repost/reply/like
use of URLs
and hashtags
meta information
(time, source)
4 Microblogging Behavior on Sina Weibo and Twitter
User modeling and analysis framework
syntactic analysis
sentiment analysis
Cultural
Analytics
semantic analysis
temporal analysis
Data Processing and User Profiling
data acquisition
multilingual NER
metadata extraction
semantic enrichment
topic/interest modeling
user profile construction
Microblogging Platforms
Semantic Web
5 Microblogging Behavior on Sina Weibo and Twitter
User profiling
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
?
6 Microblogging Behavior on Sina Weibo and Twitter
User profiling – syntactic characteristics
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
#
hashtag/URL
7 Microblogging Behavior on Sina Weibo and Twitter
User profiling – semantic characteristics
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
#
hashtag/URL
entity
8 Microblogging Behavior on Sina Weibo and Twitter
User profiling – semantic characteristics
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
#
hashtag/URL
entity
T
topic
topic:person
topic:location
topic:organization
9 Microblogging Behavior on Sina Weibo and Twitter
User profiling – sentiment characteristics
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
#
hashtag/URL
entity
T
topic
sentiment
positive
negative
neutral
10 Microblogging Behavior on Sina Weibo and Twitter
User profiling – temporal characteristics
Profile
What could go wrong? RT
@DRUGE_REPORT
Washington: White House
seeks ‘Control’ over
communication during ‘crisis’
#obama drudge.tw/LVZygj
#
hashtag/URL
entity
T
topic
sentiment
temporal information
11 Microblogging Behavior on Sina Weibo and Twitter
Analysis of users’ microblogging behavior
• Datasets
• Microblog data collected from Sina Weibo and Twitter over a period of
three months
• > 46 million micropost overall – 22m from Sina Weibo and 24m from
Twitter
• a sample of 2616 Sina Weibo users and 1200 Twitter users
• Analyze and compare user behavior on Sina Weibo and
Twitter
• on two levels (i) the entire user population and (ii) individual users
• from different angles (i) syntactic, (ii) semantic, (iii) sentiment and
(iv) temporal analysis
• relate our findings to theories about cultural stereotypes (Hofstede’s
cultural dimensions)
12 Microblogging Behavior on Sina Weibo and Twitter
Cultural model: Hofstede’s cultural dimensions
• Describes stereotypical cultural characteristics of nationalities
• Five core dimensions:
•
•
•
•
•
Power Distance (PDI)
Individualism versus Collectivism (IDV)
Masculinity versus Femininity (MAS)
Uncertainty Avoidance (UAI)
Long-Term Orientation (LTO)
• Scores are relative wrt. other
nationalities
geert-hofstede.com
13 Microblogging Behavior on Sina Weibo and Twitter
Syntactic analysis – what are the syntactical
characteristics of messages?
hashtag-Twitter
URL-Twitter
URL-Weibo
hashtag-Weibo
Hashtags and URLs are less
frequently applied on Sina
Weibo than on Twitter.
Users on Twitter are more triggered by
hashtags and URLs when propagating
information than on Sina Weibo.
14 Microblogging Behavior on Sina Weibo and Twitter
Syntactic analysis – what are the syntactical
characteristics of messages?
high collectivism
(Sina Weibo)
Hashtags and URLs are less
frequently applied on Sina
Weibo than on Twitter.
Cultural
Differences
high individualism
(Twitter)
Users on Twitter are more triggered by
hashtags and URLs when propagating
information than on Sina Weibo.
15 Microblogging Behavior on Sina Weibo and Twitter
Semantic analysis – what kind of topics are discussed?
Weibo
Twitter
The topics that users discuss on Sina Weibo are to a large
extent related to locations and persons. In contrast to Sina
Weibo, users on Twitter are talking more about
organizations (such as companies, political parties).
16 Microblogging Behavior on Sina Weibo and Twitter
Semantic analysis – what kind of topics are discussed?
high collectivism
(Sina Weibo)
Cultural
Differences
high individualism
(Twitter)
The topics that users discuss on Sina Weibo are to a large
extent related to locations and persons. In contrast to Sina
Weibo, users on Twitter are talking more about
organizations (such as companies, political parties).
17 Microblogging Behavior on Sina Weibo and Twitter
Sentiment analysis – what are the sentiment
characteristics of microposts?
Weibo
Twitter
Sina Weibo users have a stronger tendency to publish
positive messages than Twitter users.
18 Microblogging Behavior on Sina Weibo and Twitter
Combining semantic and sentiment analysis
The difference is amplified when discussing ‘people’ or
‘location’, with Sina Weibo users even more positive and
Twitter users more negative.
19 Microblogging Behavior on Sina Weibo and Twitter
Combining semantic and sentiment analysis
long-term orientation
(Sina Weibo)
Cultural
Differences
short-tem orientation
(Twitter)
The difference is amplified when discussing ‘people’ or
‘location’, with Sina Weibo users even more positive and
Twitter users more negative.
20 Microblogging Behavior on Sina Weibo and Twitter
Temporal analysis – how quickly do users propagate
information?
Weibo
Twitter
time distance =
trepost - toriginal post
Twitter users repost messages faster than
Sina Weibo users.
21 Microblogging Behavior on Sina Weibo and Twitter
Temporal analysis – how quickly do users propagate
information?
large degree of
power distance
(Sina Weibo)
Cultural
Differences
low degree of
power distance
(Twitter)
Twitter users repost messages faster than
Sina Weibo users.
22 Microblogging Behavior on Sina Weibo and Twitter
Qi Gao et al. Information Propagation Cultures on Sina Weibo and Twitter. In
Proceedings of ACM Web Science Conference 2012. Evanston, USA.
23 Microblogging Behavior on Sina Weibo and Twitter
Conclusion and future work
• What we did
• user modeling framework for culture-aware user modeling based on
microblogging data
• data-intensive analyses deliver valuable insights for multilingual and cultureaware user modeling
• Findings
• key differences between Chinese and US/Western users’ microblogging behavior
– e.g. Chinese microblogging activities are more positive and less ‘political’
• some of the differences can be explained with cultural model from social science
research – e.g. Hofstede: individualism vs. collectivism
• Future work:
• develop personalized applications that are able to adapt to the cultural factors
24 Microblogging Behavior on Sina Weibo and Twitter
Thank You!
Q&A
Qi Gao, Fabian Abel, Geert-Jan Houben, Yong Yu
[email protected]
@wisdelft
25 Microblogging Behavior on Sina Weibo and Twitter

similar documents