Cool Data Visualizations Shed Light on Chinese Microblog Censorship


Censorship in China is a pretty frustrating thing (at least for those who live here), not least because the system is so opaque. It’s never entirely clear what the standards are, or why certain things get censored while others don’t. Thankfully, researchers around the globe are digging into how censorship works here, and today we’ve come across two different articles that shed some light on how censorship works on Sina Weibo.

The first comes from Business Insider and concerns how fast Weibo censors delete critical posts. Dr. King-Wa Fu tracked a single critical Weibo post over time and found that censors deleted it within three hours:

The Business Insider headline calls this “insanely fast,” but it’s clearly not fast enough, as Dr. Fu said that within those three hours the post was exposed to tens of millions of weibo users. For more details, I suggest you check out the original article.

Meanwhile, the Nieman Journalism Lab has also dug deeper into Chinese censorship via the work of Chi-Chu Tschang, an MBA student at M.I.T. The article is quite detailed and I highly suggest you read it, but it includes some interesting data visualizations we wanted to share here, too. The first is this fascinating graph, which shows the number of weibo posts deleted over time. Onto that graph, Tschang has marked some of the biggest “sensitive” political events of this year. Do you see any correlations?

Click to view the full-size image

Tschang also put together a word cloud of what terms appeared most frequently in deleted weibo posts. Perhaps unsurprisingly, the biggest terms here (and most of the small ones) are totally innocuous — “re-posted,” “ha ha,” “good,” etc. — so there’s definitely more to the mystery of what exactly gets censored and why:

Anyway, if you’re interested I highly recommend you check out both of the original articles for more details and analysis.

[via Business Insider and Nieman Journalism Lab]

(And yes, we're serious about ethics and transparency. More information here.)

Read More