One thing to note about terrorist attacks, is they follow a power law. The vast majority of terrorist attacks are complete failures that cause barely any casualties. But the most successful ones kill thousands. It's quite possible an unusually successful terrorist attack could kill tens of millions, which would vastly overwhelm homicide statistics. E.g. if a terrorist organization obtained weapons of mass destruction. Fortunately that has lowish probability, but perhaps not low enough to make the risk greater than homicide.
Say there is a 1% chance every decade of a terrorist attack that kills 15 million people, the annual expected death toll is about 15,000, which is greater than the annual homicide deaths.
The estimated immediate casualties from another nuclear state air-burst nuking every major US city simultaneously is only like 60m. A terrorist attack would be able to achieve significantly less.
I get your general point, but what I'm going for is roughly: humans are total crap at properly estimating, predicting, and gauging the risk of crazy anomalous events like that. (I think if we weren't so bad at it, less people would play the lotto for one...)
Since I don't intend on shooting myself, the probability of me being shot is immediately a fraction of the number quoted in the article.
While I did not review the data for 2005-2013 and 2015, if the trends are similar in those years, then firearm deaths are about 1/3 (32.6% in 2014).
I live in Ohio.
If you're looking at regional differences, the ratio to firearm homicides to terrorism deaths is even higher in Ohio since there has been no major terrorist attack there, but there have been hundreds of firearm homicides over that 10 year period.
[0] - https://www.cdc.gov/nchs/data/nvsr/nvsr65/nvsr65_04.pdf table 18, firarm, homicide