Twitter as crowdsourced early warning


Posted on September 25, 2011  /  3 Comments

The idea of using information supplied by people for early warning is extremely attractive. So much so that one politically-correct person wanted us to rename our project from “last mile” to “first mile.” We didn’t because in our model it was the last mile, the end of the warning chain, and we have little tolerance for people who think the world will change simply because we rename it. But that does not stop us from thinking about the possibilities of detecting hazards through crowdsourcing. Seems quite appropriate for “unnatural” hazards of criminality as described in this report:

“Avoid Plaza Las Américas,” several people wrote, giving the location.

“There are gunmen,” wrote others, adding, “they’re not soldiers or marines, their faces are masked.”

These witness accounts have become common in Mexico over the past year, especially in violent cities where the news media have been compromised by corruption or killings.

These witness accounts have become common in Mexico over the past year, especially in violent cities where the news media have been compromised by corruption or killings. But the flurry of Twitter messages about the bodies arrived at a telling moment — on the same day that Veracruz’s State Assembly made it a crime to use Twitter and other social networks to undermine public order.

Of course, we need to think about what could go wrong. Could criminals wanting to clear an area use Twitter for their purposes?

3 Comments


  1. Reminded me of this comic: http://xkcd.com/723/ :D

  2. Good one!

  3. Not for disaster situation but @trafficsl does a similar thing.