In 2011, society was introduced to the concept of “predictive policing”; an approach to policing that would use analytics and predictive models to increase the effectiveness of the police in most communities.
On the surface, it sounds like a utopia – the use of data can only increase the safety of your home and those around you and who doesn’t want to live in safe community, right? Fast forward to 2016 and our news outlets are filled with higher crime rates in major cities like Chicago, very public shooting incidents involving citizens & law enforcement and an increasing overall mistrust of the police by the public. Whether that mistrust is appropriately placed or not is not for this article to decide. Many police officers and departments are constantly put in situations with aspects of society and the law that most average citizens can’t even begin to comprehend.
However, I do believe that there is a middle ground here. For far too long, there has been an “us versus them” mentality between the public and law enforcement (and vice-versa!). If we are to stop this approach and start to work together as a society (police officers have homes too, after all) then we need to identify some of the items that keep the two sides apart and use them to build trust and co-operation.
If predictive policing is to work, then it should be done out in the open with both police and community involvement. Make these efforts open source, let people see what goes into the models and contribute to refining those models. The second a police department implements a predictive policing approach without community involvement is the second that effort fails.
Here’s where predictive policing can play a role. As a data scientist, I believe in the power of analytics and predictive models. What I don’t believe in is using models that are hidden away from the public (“blue box”, if you will) that will in turn affect those same citizens. I strive to be an advocate for an approach that makes police and community data freely and widely available and support bridges between law enforcement and citizens that can develop models together.
My initial contribution to this effort is VegasCrimeStats.Org. By using some of the available call data from the Las Vegas Metropolitan Poloce Department, I am presenting a basic level of analytics to the general public. A primary step in lowering crime is increasing citizen awareness of the crimes around you. This web site aims to increase that awareness and will grow over time as suggestions come in from readers.
While analytics are not quite predictive, some analytical-only models (such as heat maps) are successfully used by smaller police jurisdictions that may not have the data or budget to implement a larger-scale approach with fully predictive models. A public effort has to start somewhere and I am presenting VegasCrimeStats.Org as the first part of that effort. The more data that is made available to the public, the more that citizens can be involved with (1) awareness and (2) take the next step into developing open-source predictive policing models that can be revised and used by anyone with an interest in a better community.
Successful models for an orderly community are like the cure for cancer or AIDS; much too valuable to society for them to be in the hands of a single company/city/government. Make these predictive models freely available to anyone that wants to see them, refine them and implement them.
Obviously, this is a topic that I will write about more in the future. If you have know of additional sources of crime data for the Las Vegas area – or have suggestions for other analytics that you would like to see – I would love to hear from you at crimedata [at] buriedinfo.com