Comparison of Information Spreading on Sina Weibo and Twitter Network

April 1, 2019

Our first-year PhD student, Hao Cui, presented her thesis proposal on March 26, 2019, on the comparison of information spreading on Sina Weibo, a social media platform in China, and Twitter.

Sina Weibo and Twitter are two of the major online social network sites. They share many similarities in size, structure, and influence (Figure 1).

Figure 1. Comparisons of Sina Weibo and Twitter

The comparison between Weibo and Twitter could be interesting because of the following reasons:

  1. There is no geographic overlap between the two;
  2. The functionalities between the two platforms are the same.

Compared to Twitter, Sina Weibo has received less attention from the research community even if the platforms are highly comparable.

There are several techniques to test the information diffusion in online social networks. However, she will focus mainly on such techniques adopted to model the diffusion of processes which can be based either on explanatory models (static or dynamic networks – Figure 1) or predictive models (graph or non-graph-based – Figure 2). On the other hand, the techniques adopted to identify the influential spreader will be mainly of three kinds (Figure 3): topological approaches, identification using diffusion models or using users’ features.

Figure 2. Explanatory models

Figure 3. Predictive models

Figure 4. Identification of influential spreaders

The existing dataset about the Sina Weibo network has been crawled by Tsinghua university, compressed RAR format size 8.15G, from July 28, 2012 to October 29, 2012. It includes the following features:

1. Static Following network

2. Dynamic Following network

3. Retweet behaviors with retweet content

4. Merged retweet behaviors without retweet content

5. User profile

6. Original tweet content

7. User id map

Figure 5. Data statistics

The main research objective will be to understand the dynamics and model the information spreading phenomena on Sina Weibo and find interesting differences by comparing with Twitter. Moreover, after knowing how the information spreads on Sina Weibo, It would be possible to model the ranking of hottest topics in real-time.

There are two main obstacles to overcome:

  1. Closed world assumption: to relax this assumption, one need to align users’ profile across multiple social networking sites. Observe the information diffusion among various platforms simultaneously subject to the availability of data.
  2. Cooperating and competing diffusion processes: the described studies rely on the assumption that diffusion processes are independent, i.e. each information spreads in isolation. However, Myers et al. argue that spreading processes cooperate and compete.

Blog post by Teodoro Criscione