When people enjoy themselves on making friends on Facebook, sharing news on Twitter, another influential social network called Weibo is drawing more attentions to researchers since they want to investigate what distinguished properites Weibo has and how it becomes dominant in China. However, the enormous size of Weibo which is claimed as 500 millions prevents the researchers from conducting representative analysis. Combining different sampling methods together, this study collects representative sampled data from Weibo database and gives statistic results on users' properties.

Along with the sampled data, we also collect a sub-network which is a relation network between top 224 users. On this sub-network, we apply different ranking algorithms to figure out who are the top-ten popular users, who are the top-ten influential users and who are the top-ten key persons. Furthermore, with integration of all the data we have, we detect the communities from the sample user network. Interestingly, the densities of some communities are incredibly high, based on this, we classify a big world of social network ghosts.

Degree information of Uniform IDs could be downloaded here

The first column contains 1.18 million uniform IDs, the second column represents in-degree, the third column represents out-degree, the forth represents message.

A general description of the data and the experiment is Here