Keep in mind that "gun deaths" include suicides. I'm not sure what kind of policy choice is informed by counting suicide and homicide under the same category. But it does make the gun problem seem more severe to the inattentive reader.
What this article doesn't point out is that CDC, DOJ, and FBI estimates, as well as a range of private studies and surveys, put defensive firearms usage about a conservative minimum of 300,000 instances per year, though it can range to around 1-3 million depending on the source. I would normally discard that kind of range, but the numbers have been fairly consistent in the 300,000 to 1,000,000 range. That means that for every person killed with a gun, either through homicide or suicide, there are at least 7 people who defend themselves with one.