Software to predict crime scenes

Santa Cruz, California police are testing prototype software that predicts where crimes may be committed in the next few days. The deputy chief of police thinks that it may help police patrol areas that aren't hotbeds of shady activity. Santa Clara University mathematician George Mohler developed the algorithm. From New Scientist:

28333401 Some crimes follow potentially predictable patterns. One burglary, for example, tends to trigger others nearby in the next few days, rather like aftershocks from an earthquake. In 2010, Mohler's team turned equations used to predict aftershocks into the basis for a program that uses the dates and times of reported crimes to predict when and where the "after crimes" will occur.

On average the program predicted the location and time of 25 per cent of actual burglaries that occurred on any particular day in an area of Los Angeles in 2004 and 2005, using just the data on burglaries that had occurred before that day…

Mohler and his colleagues will conduct a controlled experiment with the Los Angeles police department later this year. Officers will run the prediction algorithms as they do in Santa Cruz, but patrol only half of the locations it flags. They will then compare crime levels in the two groups.

"Cops on the trail of crimes that haven't happened" (New Scientist)

Santa Cruz Experimental Predictive Policing Software (UC Santa Cruz)

29

    1. Aren’t most crimes these algorithyms are meant to predict, crimes of opportunity? If the opportunity is there they’ll probably be committed, whether or not criminals know the methodology of police tactics.

      Now if you argued that once police presence increased in a specific area, crime would divert to another area, then yeah, that stands to reason.

  1. Opal, jasper, agate, malachite, tourmaline, beryl, porphyry, sapphire, cinnabar, turquoise, tiger’s eye, garnet, topaz, taafite, pyrite….

  2. These “prediction algorithms” sound familiar.  I remember that idea from the movie “Fail Safe”.  Here is the pertinent section from the Wikipedia entry on the “Fail Safe” movie: “Like the American attack order, the Soviet decision to initiate the jamming had been neither made nor authorized by any human being. Instead, it had been done automatically when computer algorithms, employing what the Chairman calls “their own logic”, for some reason determined that the standard American alert maneuvers might on this occasion be a real attack.”

    What could possibly go wrong?

    1. “What could possibly go wrong?”

      Cops will patrol somewhere a crime isn’t happening?  What are you imagining might happen?

      1. Well, when it was tried in San Francisco, in the 1970’s, the cops blocked off both ends of the block where crime was expected to take place, and arrested EVERYBODY on the street in between.  That worked just as well as you expect it would.  Of course, in the 1970’s, the cops were not using computer driven algorithms.  So, I am sure it will all be O.K. now.

          1. Patrolling efficiently is nice.  Here is some video from a High School where the police “predicted” that there would be some crime:

            Of course, in this case no crime was found. Next time they’ll use computers!

          2. I don’t think the cops in that video were just patrolling.

            Nobody seems to be claiming this algorithm should be used to justify any specific police actions, arrests, raids, etc. If used as described this would just be a tool for helping to distribute existing resources effectively. The cops are on the street anyway, it makes sense to try to figure out where they are most likely to be needed.

          3. The cops are on the street anyway, it makes sense to try to figure out where they are most likely to be needed.

            Except you run the risk that, when you send more cops to the Richmond, they discover more crimes and it becomes an official high-crime neighborhood. Not because there are any more crimes there, just because there were more cops assigned to look for them.

          4. Except you run the risk that, when you send more cops to the Richmond, they discover more crimes and it becomes an official high-crime neighborhood.

            But they do that now, except they base the decision to send more cops to Richmond on “hunches” or “experience” or “profiling” or stats that are at least as dubious in value as what this algorithm is likely to spit out. I don’t know that this system will work any better, but I’m very curious to see how the controlled experiment works out.

          5. That emotional footage has convinced me. Police shouldn’t try to guess where crime is more likely to occur. Instead, they should spread out uniformly all over the planet. Of course, that will require a lot more police boats.

  3. I’m just happy to see the words “controlled experiment” used in a story about new law enforcement tactics. It’s almost as if they actually want to find out if it works before adopting it as standard practice.

    1. If someone gets convicted, it’s an effective tactic.  The librul hand-wringers can worry about the minor details, like if the guy actually did it, or if it was actually illegal.

      1. Dirty cops don’t need algorithms to frame people for crimes they didn’t commit. Why would it be illegal to send more police to one neighborhood on a particular night instead of another?

      2. “If someone gets convicted, it’s an effective tactic.  The librul hand-wringers can worry about the minor details, like if the guy actually did it, or if it was actually illegal.”

        I’m endlessly fascinated that the people who should care the most about personal liberties and freedoms (conservatives) seem to care the least about them.  Then again I may be guilty of believing that conservatives actually have a semi-coherent ideology.  Beyond making sure the world is safe for corporations, I mean.

  4. Why does one burglary “trigger” others nearby? Is it just the same burglar working a neighborhood? Are there people who hear about a burglary and think it sounds fun?

  5. After spending a few weeks in Santa Cruz last year (and having my S4 broken into mid-day, right on the street near the shops we were in) I say you can pretty much lock up everyone there and only improve the planet. Miserable place (in a great location) with a miserable lot of inhabitants.

  6. They’re not arresting people based on the algorithms, they’re just sending out police to patrol areas where crimes are statistically more likely to happen. This is really something we should have started doing a long time ago, and if implemented properly, it should be a step toward reducing violent and property crimes that could also help reduce police budgets. And the fact that they’re doing a controlled experiment to test the method is pretty awesome. This is very much a good development.

  7. I wonder if they read this report from the 1970’s: Kansas City preventive patrol experiment that showed “Increasing or decreasing the level of patrol had no significant effect on resident and commercial burglaries, auto thefts, larcenies involving auto accessories, robberies, or vandalism–crimes.”.

  8. Of course it’ll work. Crime goes up whenever there is a police presence. It’s the weirdest thing, it’s as though just /being there/ allows them to catch criminals…

Comments are closed.