I was especially interested, because St. Louis is implementing this sort of “Comp-Stat” predictive analysis. Just as there’s a KCPD-UMKC partnership, and a Memphis PD-University of Memphis partnership, there’s a SLPD-UMSL partnership in this matter, mainly because there’s a SLPD-UMSL revolving door.
The main interesting takeaways for me were:
* Even software engages in racial profiling. Which means that perhaps racial profiling is a feature, not a bug. But what it tells me is that if some people think that offloading things to software so that it can do what humans can’t because humans fear racial profiling accusations are mistaken.
* The whole thing produces very little in the way of real tangible positive results. The best it can do is affect momentary blips in homicides, that are just one or two off anomalies. I mean, why would it? You’ve got software to point you to the worst of the worst in Bell Curve City, but it’s still Bell Curve City.
* It goes back to the point I made here a few days ago. Since a lot of black murders and other serious violent crime exists within a network and a circle of do-badders, and we’re constantly told not to care about it because it’s not our problem for the most part, then why bother with any of this? If we’re not allowed to care about it as individuals, we should quit caring about it publicly and officially.