As opposed to other corners of the criminal justice system (e.g. prosecutors, probation, courts), police, especially the larger agencies, tend to be much more forthcoming with data. Many of them post their vehicle and ped stop data or will send it to you if asked. Some examples include:
Washington DCLos AngelesSeattleMinneapolis... there are many many others
Sharad Goel's team at the Stanford
Open Policing Data Project has gathered a substantial amount traffic stop data. Many of the datasets are from state police, but they also have numerous individual police departments added as well.
The
Citizen Police Data Project has decades of complaint data on Chicago PD. There's a link to their github with the raw data.
Philadelphia posts detailed information on
every police shooting, including a text description of the environment, injuries, and incident outcomes. Dallas PD also post
police shooting data. Departments vary greatly in how much actual information they provide. The Los Angeles Sheriff, for example, lists a fairly
limited amount of information on all of their shootings, but if there is an injury or fatality they will post
more details, sometimes including video and the DA's investigation.
New Orleans posts data on all use of force incidents, including take-downs, hands, tasers, pointing of firearms, and shootings.
The Police Foundation's
Police Data Initiative has links to various department's data on a variety of topics including complaints, stop data, and use of force data. Several of the links are dead and in some cases the departments have not updated the data in years. However, there are still a lot of good data sources here.
The
National Archive of Criminal Justice Data (NACJD) is a a US DOJ Office of Justice Programs sponsored repository with numerous datasets on all aspects of the criminal justice system. In there you will find various datasets on race, policing, sentencing and a variety of other topics.
Although I've encountered some exceptions, I have generally found police departments to be very willing to share data, grant access to visit training facilities, offer police ride-alongs. Especially if you want to study your local police department, departments tend to be even more responsive to requests from their own community. If there's something you want to study at a particular police department, just ask. Sometimes an individual at the police department will prefer a FOIA request so that there is some formal request for the data.
------------------------------
Greg Ridgeway
Associate Professor
University of Pennsylvania
------------------------------
Original Message:
Sent: 06-09-2020 15:45
From: Roberto Rivera
Subject: Police stop data for courses in Statistics
Recent events in the U.S. have led to an important discussion about racial disparity and bias; topics that are not exempt from requiring data-driven policy making. Surely, bringing these data sets and topics into the classroom should generate much interest among students.
Police brutality data is rather scant. But there is police stop data for many locations. The Stanford Policing Project has a great set of data curated from multiple states and cities. More updated data sets may be found in open data portals of specific locations, though data wrangling may be needed (e.g. in their open portal San Diego officials share recent police stop data). Instructors may also do a public records request of the most recent police traffic stop data.
In the recently published book Principles of Managerial Statistics and Data Science, there are case studies and problems looking into whether the probability of searches during a traffic stop in San Diego and Austin depend on the driver's race. The companion website for the book shares the R code which could be applied to other jurisdictions. For a more sophisticated analysis of racial bias (based on logistic regression of contraband found during searches), you can check the R code here.
------------------------------
Roberto Rivera
Professor
------------------------------