Iraq casualty dataset

January 17, 2007 5:15 PM

Is there a well-formatted data file of weekly or daily Iraqi civilian casualties?

I need something that includes information on the cause of death like car bomb, militia gun fire, coalition forces, etc. This information exists on the Iraq Body Count and Iraq Coalition Casualties websites, but I don't want to deal with the conversion issues to get it into a spreadsheet. The closest thing I've found is this (the second tab at the bottom), but it doesn't include the cause of death.

posted by thrako to computers & internet (3 comments total)

The best information out there is the IBC. Did you check icasualties.org? Failing that, why not email them and ask for the data? Surely they've got it in .csv or something before it's posted online.
posted by B-squared at 6:04 PM on January 17, 2007

Please be aware that the websites which track Iraqi casualties are only tracking news stories. I would be surprised if 10% of all Iraqi deaths were reported in the news - there are something like 9 (nine) foreign reporters in all of Iraq, for instance.

If you're Iraqi and you die in some fashion that doesn't involve a giant car bomb within a few miles of the Green Zone (i.e., the smoke is visible from the Green Zone), guess what, you're not going to get counted. The Iraqi government has no interest in compiling a list of civilian casualties (many of which are Sunnis killed by the Shiite government). The U.S. has no interest in compiling a list of Iraqi civilian casualties (reflects badly on Shrub).

It's akin to determining how many people died in the United States by reading the New York Times, LA Times, and Washington Post. Let's see, 12 obits in the NYT, 17 in the LAT, and 14 in the WaPo - so 43 people must have died in the United States yesterday. Do you see the flaws in that methodology?

Honestly, you might as well just make up a number.
posted by jellicle at 6:31 PM on January 17, 2007

B-squared - yeah, a CSV is exactly what IBC (and now I) have.

jellicle - I'm interested in the changes in death rates, not the levels. So with any luck underreporting shouldn't bias the results too much.
posted by thrako at 4:46 PM on January 20, 2007