Angry Jesus writes:
"The Chicago Police Department is mis-applying epidemiological science (the study of entire populations) to target individuals in a real-life version of Minority Report. They have decided that it is a good idea to put people on a secret list based on a Big Data analysis of their social networks. But don't worry, it isn't racist or abusive because, Science!"
Their algorithm might actually be more accurate if it DID look at race, and gender.
There are some undeniable truths in criminality when it comes to serious crimes such as: armed robbery, rape, and murder.
1) Men are more dangerous than women. Its not that women don't kill, but when they do its usually by poisoning or what is termed a multiple offender killing *(they hire a hit man). Even adjusting for this, for those three crimes you would be hard pressed to put the odds are less than 90 times greater for a man to be involved in those crimes than a woman.
2) Black men are more dangerous than White men. Hispanic men are more dangerous than White men. With the exception of gang violence, Whites men are more dangerous than Asian men.
3) Violent crime is primarily committed by those between the ages of 15-34. At a certain point, people mostly age out of it.
4) Being involved in a gang increases your likelihood of being involved in a homicide both as a perpetrator and as a victim.
5) Being politically correct is a waste of law enforcement resources. I'll be called racist for this simple truth: The single best predictor of the violent crime of an area is the percentage of blacks and hispanics that live in the area. If you want to find your hot spots, demographics will do that for you better than any other single tool. Combining tools should prove even better.
6) Add in gang affiliation, people arrested at known drug corners, affiliations with other known gang members, race, sex, age, and prior law enforcement contact -- and I would be willing to bet you could make a pretty darn good profile of people thousands of times to even hundreds of thousands of times more likely to be involved in certain offenses than the average person in a community.
7) If its okay to profile and pay preemptive visits and do tracking for those people that are sex offenders, why not for the gang unit? Or the robbery unit? This isn't targeting people who have no contact with the criminal justice system. Its targeting people who have already become known to the justice system, and have a whole host of other unsavory connections to predictors of violent crime.
So long as its merely used as an investigation tool, and for helping to decide the best deployment of resources I don't have a problem with it. Analytic data on such items as auto theft has helped the St Louis County police department greatly reduce auto theft simply because they deployed their resources and bait cars to those areas, and within the normal traffic routes of those suspected of being involved in chop shop operations. It also allowed them to deploy the "most wanted" cars as well.
Now if this technique were to add restrictions on these citizens that prevented them from traveling or something along those lines -- I would take exception to it. That does not appear to be the case.
I doubt that. Can you cite any data that takes into account other relevant factors?
There's a correlation between being black and being poor. And there's a correlation between being poor and being violent. Thus there's certainly a correlation between being black and being violent. But that doesn't make black men more dangerous than white men. Indeed poor white men will, on average, be more dangerous than wealthy black men.
You'll also find that tax evasion is more common for white men than for black men. Does that mean white men are more greedy than black men?