Predict Crime Like You Predict The Weather
from the tomorrow-should-be-high-in-vandalism,-but-a-low-in-shoplifting dept
Not quite to the level of “Minority Report”, but researchers are working on ways to predict crime in any given area. The work is actually similar to predicting the weather – taking different bits of trending information and using it to predict future results. So far, they say that within a given police beat, they can predict crime levels (on a monthly basis) accurately 80% of the time. The best leading indicators are apparently “soft crime” increases – such as “mischief, disorderly conduct, and trespassing”. As they point out, it seems pretty obvious, but this is the first time anyone is doing anything with it. Of course, the other issue is that (unlike the weather), their predictions can impact actions (send more police into a predicted high crime zone), making the end results seem less accurate.
Comments on “Predict Crime Like You Predict The Weather”
Predicting crime
“As they point out, it seems pretty obvious, but this is the first time anyone is doing anything with it.”
Au contraire. Way back when I was in high school — I graduated in 1983 — my class got a visit from a police officer who was pissed off that the department was using (no doubt rudimentary) statistical analysis to try to predict where they should send squad cars on patrol. This was in Eugene, Oregon — not exactly a center of computer innovation — so it must have been a practice somewhere else as well. Or maybe I’m wrong, and Eugene was the home of some lone, unappreciated criminology genius who was 20 years ahead of his time.
PD vs. FD Bloodsport
What if such a program can show an increase in looting incidents whenever fires break out? Will cops show up at burning buildings and arrest pilfering firemen?
I’ve seen those guys fight before, both types are very powerful men, and it’s pretty scary being near them. They smash right through glass doors and keep fighting, like Matrix or something.