>> Actual homicide is just the tip of the iceberg, of course, but if this doesn't qualify as "significantly and often" then I'm not sure what does.
Mainly because police departments do not keep these kinds of statistics, so the data is horribly incomplete. And as anybody knows, you can massage the data to show anything you want when you don't have the complete picture:
Officials with the Justice Department keep no comprehensive database or record of police shootings, instead allowing the nation’s more than 17,000 law enforcement agencies to self-report officer-involved shootings as part of the FBI’s annual data on “justifiable homicides” by law enforcement.
That number – which only includes self-reported information from about 750 law enforcement agencies – hovers around 400 “justifiable homicides” by police officers each year. The DOJ’s Bureau of Justice Statistics also tracks “arrest-related deaths.” But the department stopped releasing those numbers after 2009, because, like the FBI data, they were widely regarded as unreliable.
source: https://www.washingtonpost.com/news/post-nation/wp/2014/09/0...
The other problem with your data is that there is no firm evidence how each of these scenario's played out. Did the person being shot pull a firearm? What about brandishing a weapon or charging the officer, or attempting to assault the officer? We don't know because this kind of evidence is never taken into account. All people see is a cop who killed a citizen, nobody knows the exact circumstances around every single one of these instances - which makes ALL the data that much more suspect.
So yeah, when you look at the raw numbers, they look bad, but when you add in the circumstances by which these played out, you'd probably find out a lot of these were justified - which is why cops tend to get a lot of leeway in these kinds of scenarios, and are rarely charged.