Censored Sina Weibo posts translated into English by HKU project
Some posts deleted by censors from mainland microblogging site Sina Weibo have been recovered and are now available in English.
WeiboScope, a program that can recover deleted posts - including those removed by government censors - has collected 200 million posts since its launch at the University of Hong Kong (HKU) in 2011.
And now a team of data journalism students from the university's journalism school have developed a search engine that makes some of those posts accessible to English-language users.
The students call their translation project "WeiboSuite" and hope it will become a valuable tool for journalists.
"The Chinese internet is a closed book for most English-language speakers," said Henry Williams, a student from Britain. "By creating WeiboSuite, we want to make a very important part of it much more accessible to journalists across the world."
More than 300 million mainlanders use Sina Weibo to communicate with each other. Some have used the platform to expose corrupt officials and spread information that would otherwise be censored or played down.
WeiboSuite shows posts in their Chinese format followed by an English translation, but is restricted to searches of 1,000 of the most recently deleted posts.
"It works well because journalists normally only need current weibos, but we would like to expand service," Williams said.
James Griffiths, an editor at Shanghaiist.com a lifestyle blog, said he began using the search engine a few weeks ago. "I hope that WeiboSuite can become a genuine alternative to Weibo.com for journalists and academics," he said. "Weibo is invaluable for getting an impression of the mood among [internet users]."
Williams and classmates Jacky Wong and Tony Yoo, who developed WeiboSuite, earned scholarships from Google to study data journalism.
WeiboScope has been developed by HKU's Journalism and Media Studies Centre. Dr Fu King-wa, the program's lead developer, said research this month based on WeiboScope data showed censors had been increasingly successful in stopping users posting sensitive content since the government began requiring microbloggers to register accounts using their real identities in March last year.
"By comparing the activity of a group of users for three months before and after the real-name registration policy, we observed an overall drop in usage, and hypothesise that the policy had a chilling effect on some people," Fu said.