[kfor]

I head up the scientific computing team at IHME (the institute the article talks about), happy to answer questions.

I started out doing a fellowship [0] there, it was a great first job.

[0] http://www.healthdata.org/get-involved/fellowships

[aamar]

Thanks for the opportunity. This is an interesting project, and the linked graphs are intriguing:

http://vizhub.healthdata.org/gbd-compare/#

Questions: How are DALYs attributed to various risk factors? Assuming this is extracted from the data itself, what model do you use to isolate causation from correlation?