What about it makes it disproportional? Blacks make up a lesser percent of the population, but if a black man goes his entire life without seeing a policeman, his chances of dying to one go down dramatically.
What if we did something like control for number of police encounters, to see if there's some kind of racism involved. We could use arrests as a normalising factor, a stand-in for police encounters.
Else you run into the trouble where whites are disproportionately killed by aircraft and having buildings collapse on them, and then concluding that aircraft and buildings are racist. Or… it is more likely for white people to be in giant buildings, or on airplanes. Or in boating accidents. If you're going to judge the risk to someone of an event, you have to control by the number of events, not by population, as there is no guarantee that it is proportionate otherwise.
Problem is when you do, you find out the disparity vanishes. Mostly, anyway. Whites, per arrest, have roughly the same chance of dying. Ergo the problem is not with police brutality, violence, or how they handle arrests, but in the rate at which they arrest people.
If you dig into that, you might find that they arrest people based on race more often. Racist arrest rates, for example.
How would you measure this? you would need a statistic for reported crime+suspect which then led to arrest. But reported crimes are full of biases, but really what else could we collect.
Really what the problem is, at heart, is that people who commit more crimes encounter police more often. This increases their chance of being dealt with lethally.
Another option for studying this could be to collect and monitor body cam footage. You categorise each arrest and collect what happens. For example; violent resistance, resisting arrest, etc. Although this isn't done either because it turns out that blacks resist arrest more often.
And yet per arrest the difference is minimal. The police being bad is a myth. Criminals being bad, oh we forget about THAT.