Assuming for a second that the question was asked with legitimate intent: That's the line of reasoning that gets a lot of innocent people treated very very badly based on their socioeconomic status, color of their skin, religion, you name it.

Think of it this way: In 1,000,000 people of race blue, the crime rate is 0.01%. So of those million people: 10 are criminals. In 1,000,000 people of race purple, the crime rate is 0.03%. So 30 are criminals.

All kinds of horrible sounding statistics here: purple people are 3x more likely to be criminal. Purple people neighborhoods have 3x more crime. Purple people 3x more likely to be murderer, drug user, etc.

All of that misleads about the situation: for the most part both purple and blue people are law abiding and any given person you talk to is likely not a criminal.

Now there is a second order problem. (The one that the article is referencing). The truth is that among blue and purple the rate of criminals is much higher. Pretend it's +0.3 for each. So 0.04 and 0.06. But the enforcement of laws at each point in the system are subject to discretion of the enforcer. Does the cop pull you over? Does the cop let you off? Does the prosecutor indict? Does the system give you bail?

Because of focus on "where would they find the most amount of criminal activity?" we are now creating a self fulfilling prophecy where laws are selectively more greatly enforced against purple.