One Step Closer To Minority Report: We Know Where Crime Will Happen
from the time-to-go-somewhere-else dept
Earlier this year, we had a story that claimed some police had used some data mining software to predict a mugging, though the details on the story were incredibly vague — leading to plenty of questions about just how legitimate the story really was. However, police are clearly interested in this Minority Report-style software. Police in Richmond, Virginia, are now getting their very own data mining solution to help them predict crime, though it’s less about predicting single instances of crime, but noting general patterns and suggesting where police might want to go. It’s not a bad idea to use the data in this way, but you always get concerned that when people become too reliant on the data, mistakes tend to happen.
Comments on “One Step Closer To Minority Report: We Know Where Crime Will Happen”
Even more jobs for statisticians
There are already plenty of juicy pickings in the pharmaceutical or insurance industries, but sounds like now we have even more opportunities. We’re not quite clairvoyants, though statistics can make amazing predictions, and find patterns that defy common sense.
Anybody willing to take the prerequisite courses can go to grad school for stats — you don’t have to have a math degree. The programs are in various names like statistics, biostatistics, actuarian sciences, industrial optimization, survey methodology, etc.
Re: Even more jobs for statisticians
It’s not even statistics. Some data, along with some geocoding in a program like ESRI’s ArcMAP/ArcGIS package can yield the same sort of predictions. Police departments nationwide are using geographic mapping software packages to help hone in on these trends.
Data and a little geocoding can reveal all types of trends.
Re: Re: Even more jobs for statisticians
The people who program the packages have to know statistics. As the software becomes more sophisticated, it will take statistical expertise to correctly interpret the results also. Even in academic circles, social sciences like criminology are increasingly accept the need for better statistical expertise to make sense of data, or to validate theories.
Re: Re: Re: Even more jobs for statisticians
I agree, I’m not knocking the validity of statistics. I’m just pointing that the average user of these packages doesn’t have to be a statistician. Some of us have degrees in Geography. Some (police officers who are in charge of this data) can’t even decipher a standard deviation from a median from a hole in the wall.
Those programmers have to have some statistical knowledge (or at least a stats text book handy), but the average user doesn’t.
Re: Re: Re:2 Even more jobs for statisticians
To have software that makes amazingly accurate predictions requires sophisticated statistical models to match. Having an understanding of GLMs, knowing the distinction between a Poisson process and a Poisson distribution, overdispersion, curved exponential families, Martingale theory, asymptotic statistics vs. Almost Sure convergence vs. Quasi-Likelihood methods — knowing all that requires more than just a programmer with a stat textbook handy.
Everything has some form of pattern.
This topic just amuses me. If they could why didn’t they have/send warning ahead against major disaters or would that leak their precious information.
No Subject Given
47.3% of all stats are made up on the spot.
Re: No Subject Given
A statistician’s job is to reduce that number.
Re: No Subject Given
so you are telling me that these facts are just made up
Re: No Subject Given
=p
Thanks for the afternoon chuckle.
The Usual Suspects
O: (The senior police officer?s voice has a cutting edge.) Come in, shut the door. Now then, Savage, I want to talk to you about some charges that you have been bringing in lately. I think that perhaps you?re being a little overzealous [[a little too keen]]… Savage, why do you keep arresting this man?
S: He?s a villain, sir.
O: A villain …
S: And a jailbird.
O: (exploding) I know he?s a jailbird, Savage. He?s down in the cells now. We are holding him on a charge of possession of curly black hair and thick lips.
S: Well, … well, well, well there you are, sir.
O: You arrested him, savage!
S: (stupidly pleased) Thank you, sir.
O: Savage, would I be correct in assuming that Mr Cudoogo is a coloured gentleman?
S: Well, I can?t say I?ve ever noticed, sir.
Constable Savage script