Is this a fallacy of division?

The main problem I see is confusing causality with correlation. We are concluding a cause (racism) based on a correlation (African-American individuals are more likely to have excessive forced used against them). If the fact is that African-American individuals are more likely to have excessive forced used against them, we need to find out why—we simply cannot assert that it is racism. I am speaking as a logician here, not a social scientist (i.e, this example does not reflect my academic opinion on race and violence). To remove any biases consider the following:

Labrador Retriever puppies are more likely than other breeds to be held by children.
Poppy, the Labrador Retriever who was held by Johnny was held because he was a lab.

It might be the case that Johnny loves labs, or that Poppy came over to Johnny, or many other reasons.

I don't think this is an example of the fallacy of division.

The Fallacy of Division suggests that, since trait X is true for the group Y, it must also be true for the parts of that group. 

So "more likely to have excessive force used against them and killed by police" (X) is assumed true for African-Americans (Y). 

Following your logic, one could that a given African-American facing disproportionate use of force by white officers could be racially-motivated. This is not the Fallacy of Division as you are only suggesting, not making a definitive statement.

However, the claim is slightly different - first you talk about excessive force with the race of the officer not mentioned, then you talk about disproportionate force by white officers. By 'disproportionate' I'm guessing you're referring to 'out of proportion' rather than 'statistically overrepresented', in which case, there's no problem.

I don't think so. If you said that the use of force is absolutely due to racism, that would be fallacious. But if you say "likely", you are saying that based on historical data the probability that the action is based on race is high. There can be different opinions - someone else might argue that the probability is low because AAs commit crime at a higher rate. Neither is a fallacious argument, IMHO, they are just coming to different conclusions based on the weight they put on different contributing factors.