Calling arrest data "already suspect" is begging the question, but in any case my view was not informed by arrest data in the first place.
For consideration... the rate of arrests isn't necessarily indicative of the rate of crimes committed.
As an analogy, red vehicles get speeding tickets more often than any other color. If that's the only piece of information you have, you might assume that the drivers of red vehicles speed more often than the drivers of other colored vehicles. Turns out this isn't the case though. In actuality, red vehicles are more noticeable, they stand out more, so they catch the attention of cops. Drivers of all colors of cars speed at roughly the same rate, but the red cars get noticed and pulled over a lot more often, which results in a higher rate of ticketing for red vehicles.
There's a similar dynamic when it comes to arrests and convictions for black people. They get pulled over more often than white people do, they get stopped and come under scrutiny more often than white people do, and they're given the benefit of the doubt less often than white people. So for roughly equivalent crimes, black people are more likely to be arrested, charged, and convicted... while white people are more likely to get a warning.
By the way, women get far more warnings than men do for equivalent misbehavior. Women get the benefit of the doubt very regularly. Women commit violent crimes at a lower per capita rate than men do, but even with that being considered, their arrests are less frequent and their sentences are lower. Women commit nonviolent and petty crimes at about the same rate as men, but are arrested and sentences quite a bit less often than men.
Justice in the US isn't blind.