Blogs

From The Staff

The Mathematics of Crime

A discussion with two of the team members who helped create the “predictive policing” computer program about the mathematics behind it.

June 16, 2016

From The Staff Mathematics Technology Human Ecology

Ad Right

The "Predictive-Policing" computer program allows law enforcement departments to tailor where police officers are placed in order to better prevent crime. UCLA mathematician Andrea Bertozzi—a Sigma Xi Distinguished Lecturer—and social scientist Jeffrey Brantingham discuss the thinking behind the program their team created.

A Storify of the discussion is below.

The Mathematics of Crime

The "Predictive-Policing" computer program allows law enforcement departments to tailor where police officers are placed in order to better prevent crime. UCLA mathematician Andrea Bertozzi and social scientist Jeffrey Brantingham discuss the thinking behind the program their team created.


  1. Not just theory -- in practice -- the "Predictive Policing" computer program is in use by the Los Angeles Police Department, among many others.

  2. Mathematically predicting #crime on short time scales nearly (if not) impossible. Focus instead is on "where" not "when." #AmSciGHO
  3. Bertozzi says mathematical models of type used to #predict #crime also work for predicting email traffic and IED attacks. #AmSciGHO
  4. Gang tags with spray paint -- like "broken windows" effect -- could be quite useful for modeling and predicting #crime. #AmSciGHO
  5. Jeff Brantingham says #police have been going through a #data #revolution as part of managing their activities. #AmSciGHO
  6. Most police departments have data on 100s of events that can be used for predictive models -J Brantingham  #AmSciGHO https://t.co/1m15fv0YbN
    Most police departments have data on 100s of events that can be used for predictive models -J Brantingham #AmSciGHO pic.twitter.com/1m15fv0YbN
  7. Like the spatial scales for #earthquakes, think of the #spatial scales for #crime, but with much smaller radii (~ a mile or so) #AmSciGHO
  8. We're much less focused on the individuals committing the crime as we are on the area as a target of the crime, -A. Bertozzi #AmSciGHO
  9. "It's not just about crime; it's a property of human behavior" Brantingham says modeling could work also on where people eat out #AmSciGHO
  10. Many reasons why mathematically impossible w/ these techniques (using available data) to predict "who" will commit crimes #AmSciGHO
  11. There are as many theories about how #poverty drives #crime as there are #sociologists in the world -- @UCLA's Jeff Brantingham #AmSciGHO
  12. "You want a #crime (#violent and non-violent) #model with the fewest parameters possible" -- Bertozzi & Brantingham @UCLA #AmSciGHO
  13. Concerning #privacy and #civil #liberties, types of #data (what/where/when) used is 90-95% reported by the #public #AmSciGHO
  14. "We assume the #criminals are identical... in our models" making the mathematics easier, says Bertozzi #AmSciGHO #preventing not #profiling
  15. The first use of these #predicting #policing programs started in Santa Cruz (CA), says Bertozzi, led to 27% decline in crime. #AmSciGHO
  16. Today's #AmSciGHO guest Jeff Brantingham is @UCLA professor of anthropology and chief of research @PredPol  http://www.predpol.com/ 
  17. Santa Cruz results --> 21-month-long program in #LA. In that first month, another 200 police departments heard about it, wanted in #AmSciGHO
  18. Initial deployment of #research took 6 or 7 years, but now it's a service for each new #police department. Fully automated. #AmSciGHO
  19. Human #crime #analyst is looking at "hotspot" maps -- predictive algorithms can look at complete #history and so be more accurate #AmSciGHO
  20. Paradox in #criminology: #hotspot #policing --> either crime disappeared or displaced. #Math helped explain which happened & why #AmSciGHO
  21. These methods are for delivering short-term predictions, not solving the root causes of #crime. #AmSciGHO
  22. "You can solve problems that are much bigger than you could on your own."–Dist Lecturer A Bertozzi on interdisciplinary research #AmSciGHO
  23. We do this #collaboration #multidisciplinary work because we can solve problems that are bigger than one can just do on one's own #AmSciGHO
  24. "#Social scientists have to... compromise about the way they think about their #research" for #quant #progress @UCLA's Brantingham #AmSciGHO
  25. Social science can explode by embracing machine learning, but not all disciplines have training to make that handshake –A Bertozzi #AmSciGHO
  26. Brantingham did post-doc work @sfiscience prior to getting started on this #predictive #policing project with Bertozzi @UCLA #AmSciGHO
  27. That concludes the event and @AmSciMag's live tweeting of The #Math of #Crime. Find the #video archived here:  http://bit.ly/27TU66r 

American Scientist Comments and Discussion

To discuss our articles or comment on them, please share them and tag American Scientist on social media platforms. Here are links to our profiles on Twitter, Facebook, and LinkedIn.

If we re-share your post, we will moderate comments/discussion following our comments policy.

×

AMSCI ICON NAVIGATION:

  • Navigation Menu
  • Help
  • My AmSci
  • Select Options (not present on all pages)

Click "American Scientist" to access home page