UPDATE: Internet traffic has been flowing in and out of Bahrain during the entire protest. Many people in Bahrain complain about slow and limited access. This may be due to a combination of censorship and high volumes, but the Internet has not been taken down at any point so far in a Egyptian style.
The NYT today reported:
"As protests have erupted in Bahrain over the last several days, the government has severely restricted the access of its citizens to the Internet, new data from an organization that monitors Internet traffic strongly suggests."
We partially confirm this with our own tool this morning. Even before reading the report we could tell that geo-tagged tweeting from key areas in Bahrain had collapsed. Our tool use exponential decay to model how a tweet looses significance over time, so a sudden unplugging of the Internet would not result in an immediate collapse, be a rapid fall in value.
Given that yesterday for high schools around Sulmaniya Hospital yesterday of 50 on our scale of 100 this mornings score of 3 would indicate a major interruption of service. Track this site.
But this is not the complete story.
We are seeing moderate levels of tweeting with geo-locations for Manama, specially Sh Abdulla Ave, Manāma, Bahrain Track this site in real time
We are also seeing significant fluctuations across location and time in Manāma this morning. All scores as significantly suppressed from highs of yesterday but there still are certain locations with modest tweeting.
Now it is important to say that our tool tracks the geo-locations assigned to tweets via the Twitter API. It is possible that false geo-tagged tweets are being generated by the government to make it look as though the Internet is still happening. We strongly believe the Bahrain government has been conducting a campaign of disinformation on twitter for several days.
It is also possible that people are simply retweeting posts. Retweets carry the original geo-tag so we will have to watch over time to be sure we are not seeing the effects of retweeting. We included this feature in our measurement because we wanted to see the overall impact of a tweet from a location on the entire world, rather than counting heads. We are more interested in the overall systems impact of tweeting at a given location. During the Arab revolts this has lead to us seeing elevated tweet levels for specific locations. But it may also meant hat a high score is only the echo effect of retweets over time.
Update: But this story is moving very fast. We have just noticed a rapid rise in the level of tweets coming for Pearl Roundabout. We can not confirm with our tool that this is actual tweeting being done on mobile devices and laptops in area. Track levels here.