Real-Time Crime Forecasting Challenge

The Real-Time Crime Forecasting Challenge sought to harness the advances in data science to address the challenges of crime and justice. It encouraged data scientists across all scientific disciplines to foster innovation in forecasting methods. The goal was to develop algorithms that advance place-based crime forecasting through the use of data from one police jurisdiction.

The Challenge had three main aims:

  1. Harness data science advances in other fields to crime forecasting.
  2. Encourage scientists from all fields to consider the challenges of crime and justice.
  3. Conduct the most comprehensive comparative analysis of crime forecasting software and algorithms to date.

Specifically, the Challenge tested how effectively and efficiently contestants’ crime forecasting algorithms could forecast police calls-for-service in four crime categories in Portland, Oregon,[1] for five forecast periods. View the challenge posting for additional details.

Challenge Winners

The scores are in and we have our winners! Winners were selected from submissions by five students, forty-two small teams/businesses, and fifteen large business.

Download the complete leaderboard (xlsx, 56 KB)! The leaderboard includes the names and scores for first, second, and third place submissions for every category, crime type, time frame, and score type (PAI and PEI*).

Download the winning submission files for students (zip, 29 MB), small teams/businesses (zip, 47.5 MB), and large businesses (zip, 50.8 MB). You also may download the Visual Basic macros that we used to determine eligibility and scores for each submission.

Winners by Contestant Type and Category
(See the detailed list of winners by category and timeframe)
Winner
Contestant Type No. of Victories
Warren
Student
18^
YurickStudent16^
LedrayStudent 4^
ScarboroughStudent2^
TAMERZONESmall Team/ Business
10
IntuidexSmall Team/ Business 7^^
BATES ANALYTICSSmall Team/ Business 5
TADICSmall Team/ Business 3
pennaikenSmall Team/ Business 3
ANDY_NIJSmall Team/ Business 2
WILLIAM HERLANDSSmall Team/ Business 2
KUBQR1Small Team/ Business 2
STEVEN YURICKSmall Team/Business2
MURRAY MIRONSmall Team/ Business 2^^
DYLAN FITZPATRICKSmall Team/ Business 2^^
PTL
Small Team/Business 2^^
CCC
Small Team/Business 1
​AltMaps
​Small Team/Business ​1^^
​CUDY
​Small Team/Business ​1^^
​GRANTHAM
Small Team/Business ​​1^^
​KOONTZ
Small Team/Business ​​1^^
​RAPTOR
​Small Team/Business 1^^
TEAM Kernel GlitchesLarge Business 9
PASDALarge Business 9
CodilimeLarge Business 8
GRIERLarge Business 5
DataikuIncLarge Business 2
JeremyHeffnerLarge Business 2
IMPAQLarge Business 2
GARAN TANALYTICSLarge Business 1
Conduent Public Safety SolutionsLarge Business 1
MARUAN ALSHEDIVATLarge Business 1

^ Victory totals in the student category include four-way tie in both PAI and PEI* scores for the one- and two-week burglary forecasts.

^^ Multiple Small Team/Business entrants tied PEI* score in the one week forecast for burglary.

Note

[note 1] The Portland Police Bureau provided five years of calls-for-service data as a training data set. Contestants were not limited to using the CFS data, but could also use any other data set of their choosing.

The data from the Portland Police Bureau is public and as such may be used by individuals/entities for secondary data analysis for non-Challenge related topics.

Download the data.

Date Modified: October 30, 2017